/[pcre]/code/trunk/ChangeLog
ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 587 - (show annotations) (download)
Fri Jan 14 19:01:25 2011 UTC (3 years, 6 months ago) by ph10
File size: 197996 byte(s)
Fix -M bugs in pcregrep

1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.12 12-Jan-2011
5 ------------------------
6
7 1. Fixed some typos in the markup of the man pages, and wrote a script that
8 checks for such things as part of the documentation building process.
9
10 2. On a big-endian 64-bit system, pcregrep did not correctly process the
11 --match-limit and --recursion-limit options (added for 8.11). In
12 particular, this made one of the standard tests crash. (The integer value
13 went into the wrong half of a long int.)
14
15 3. If the --colour option was given to pcregrep with -v (invert match), it
16 did strange things, either producing crazy output, or crashing. It should,
17 of course, ignore a request for colour when reporting lines that do not
18 match.
19
20 4. Another pcregrep bug caused similar problems if --colour was specified with
21 -M (multiline) and the pattern match finished with a line ending.
22
23 5. In pcregrep, when a pattern that ended with a literal newline sequence was
24 matched in multiline mode, the following line was shown as part of the
25 match. This seems wrong, so I have changed it.
26
27 6. If pcregrep was compiled under Windows, there was a reference to the
28 function pcregrep_exit() before it was defined. I am assuming this was
29 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
30 reported by a user. I've moved the definition above the reference.
31
32
33 Version 8.11 10-Dec-2010
34 ------------------------
35
36 1. (*THEN) was not working properly if there were untried alternatives prior
37 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
38 backtracked to try for "b" instead of moving to the next alternative branch
39 at the same level (in this case, to look for "c"). The Perl documentation
40 is clear that when (*THEN) is backtracked onto, it goes to the "next
41 alternative in the innermost enclosing group".
42
43 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
44 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
45 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
46 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
47 (*THEN).
48
49 3. If \s appeared in a character class, it removed the VT character from
50 the class, even if it had been included by some previous item, for example
51 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
52 of \s, but is part of the POSIX "space" class.)
53
54 4. A partial match never returns an empty string (because you can always
55 match an empty string at the end of the subject); however the checking for
56 an empty string was starting at the "start of match" point. This has been
57 changed to the "earliest inspected character" point, because the returned
58 data for a partial match starts at this character. This means that, for
59 example, /(?<=abc)def/ gives a partial match for the subject "abc"
60 (previously it gave "no match").
61
62 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
63 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
64 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
65 has an implication that the given string is incomplete (because a partial
66 match is preferred over a full match). For this reason, these items now
67 give a partial match in this situation. [Aside: previously, the one case
68 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
69 match rather than a full match, which was wrong by the old rules, but is
70 now correct.]
71
72 6. There was a bug in the handling of #-introduced comments, recognized when
73 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
74 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
75 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
76 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
77 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
78 places in pcre_compile().
79
80 7. Related to (6) above, when pcre_compile() was skipping #-introduced
81 comments when looking ahead for named forward references to subpatterns,
82 the only newline sequence it recognized was NL. It now handles newlines
83 according to the set newline convention.
84
85 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
86 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
87 cater for a lack of strerror(). These oversights have been fixed.
88
89 9. Added --match-limit and --recursion-limit to pcregrep.
90
91 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
92
93 11. When the -o option was used, pcregrep was setting a return code of 1, even
94 when matches were found, and --line-buffered was not being honoured.
95
96 12. Added an optional parentheses number to the -o and --only-matching options
97 of pcregrep.
98
99 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
100 can match an empty string. The code to do it in pcretest and pcredemo
101 needed fixing:
102
103 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
104 only one byte after an empty string match just before CRLF (this case
105 just got forgotten; "any" and "anycrlf" were OK).
106
107 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
108 mode when an empty string match preceded an ASCII character followed by
109 a non-ASCII character. (The code for advancing by one character rather
110 than one byte was nonsense.)
111
112 (c) The pcredemo.c sample program did not have any code at all to handle
113 the cases when CRLF is a valid newline sequence.
114
115 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
116 as a starting offset was within the subject string. There is now a new
117 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
118 negative or greater than the length of the string. In order to test this,
119 pcretest is extended to allow the setting of negative starting offsets.
120
121 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
122 starting offset points to the beginning of a UTF-8 character was
123 unnecessarily clumsy. I tidied it up.
124
125 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
126 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
127
128 17. Nobody had reported that the --include_dir option, which was added in
129 release 7.7 should have been called --include-dir (hyphen, not underscore)
130 for compatibility with GNU grep. I have changed it to --include-dir, but
131 left --include_dir as an undocumented synonym, and the same for
132 --exclude-dir, though that is not available in GNU grep, at least as of
133 release 2.5.4.
134
135 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
136 characters from a string of bytes) have been redefined so as not to use
137 loops, in order to improve performance in some environments. At the same
138 time, I abstracted some of the common code into auxiliary macros to save
139 repetition (this should not affect the compiled code).
140
141 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
142 compile-time error is now given if \c is not followed by an ASCII
143 character, that is, a byte less than 128. (In EBCDIC mode, the code is
144 different, and any byte value is allowed.)
145
146 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
147 START_OPTIMIZE option, which is now allowed at compile time - but just
148 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
149 to pcregrep and other applications that have no direct access to PCRE
150 options. The new /Y option in pcretest sets this option when calling
151 pcre_compile().
152
153 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
154 back references. Groups containing recursive back references were forced to
155 be atomic by that change, but in the case of named groups, the amount of
156 memory required was incorrectly computed, leading to "Failed: internal
157 error: code overflow". This has been fixed.
158
159 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
160 pcretest.c, to avoid build problems in some Borland environments.
161
162
163 Version 8.10 25-Jun-2010
164 ------------------------
165
166 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
167 THEN.
168
169 2. (*ACCEPT) was not working when inside an atomic group.
170
171 3. Inside a character class, \B is treated as a literal by default, but
172 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
173 causes the error). The code is unchanged, but I tidied the documentation.
174
175 4. Inside a character class, PCRE always treated \R and \X as literals,
176 whereas Perl faults them if its -w option is set. I have changed PCRE so
177 that it faults them when PCRE_EXTRA is set.
178
179 5. Added support for \N, which always matches any character other than
180 newline. (It is the same as "." when PCRE_DOTALL is not set.)
181
182 6. When compiling pcregrep with newer versions of gcc which may have
183 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
184 declared with attribute warn_unused_result" were given. Just casting the
185 result to (void) does not stop the warnings; a more elaborate fudge is
186 needed. I've used a macro to implement this.
187
188 7. Minor change to pcretest.c to avoid a compiler warning.
189
190 8. Added four artifical Unicode properties to help with an option to make
191 \s etc use properties (see next item). The new properties are: Xan
192 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
193
194 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
195 use Unicode properties. (*UCP) at the start of a pattern can be used to set
196 this option. Modified pcretest to add /W to test this facility. Added
197 REG_UCP to make it available via the POSIX interface.
198
199 10. Added --line-buffered to pcregrep.
200
201 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
202 studied, and the match started with a letter with a code point greater than
203 127 whose first byte was different to the first byte of the other case of
204 the letter, the other case of this starting letter was not recognized
205 (#976).
206
207 12. If a pattern that was studied started with a repeated Unicode property
208 test, for example, \p{Nd}+, there was the theoretical possibility of
209 setting up an incorrect bitmap of starting bytes, but fortunately it could
210 not have actually happened in practice until change 8 above was made (it
211 added property types that matched character-matching opcodes).
212
213 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
214 possible starting bytes for non-anchored patterns.
215
216 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
217 \R, and also a number of cases that involve Unicode properties, both
218 explicit and implicit when PCRE_UCP is set.
219
220 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
221 input, it could crash or give wrong results if characters with values
222 greater than 0xc0 were present in the subject string. (Detail: it assumed
223 UTF-8 input when processing these items.)
224
225 16. Added a lot of (int) casts to avoid compiler warnings in systems where
226 size_t is 64-bit (#991).
227
228 17. Added a check for running out of memory when PCRE is compiled with
229 --disable-stack-for-recursion (#990).
230
231 18. If the last data line in a file for pcretest does not have a newline on
232 the end, a newline was missing in the output.
233
234 19. The default pcre_chartables.c file recognizes only ASCII characters (values
235 less than 128) in its various bitmaps. However, there is a facility for
236 generating tables according to the current locale when PCRE is compiled. It
237 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
238 characters, are recognized by isspace() and therefore were getting set in
239 these tables, and indeed these tables seem to approximate to ISO 8859. This
240 caused a problem in UTF-8 mode when pcre_study() was used to create a list
241 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
242 which of course cannot start UTF-8 characters. I have changed the code so
243 that only real ASCII characters (less than 128) and the correct starting
244 bytes for UTF-8 encodings are set for characters greater than 127 when in
245 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
246 altogether.)
247
248 20. Added the /T option to pcretest so as to be able to run tests with non-
249 standard character tables, thus making it possible to include the tests
250 used for 19 above in the standard set of tests.
251
252 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
253 reference to a subpattern the other side of a comment that contains an
254 opening parenthesis caused either an internal compiling error, or a
255 reference to the wrong subpattern.
256
257
258 Version 8.02 19-Mar-2010
259 ------------------------
260
261 1. The Unicode data tables have been updated to Unicode 5.2.0.
262
263 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
264 configured.
265
266 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
267 original author of that file, following a query about its status.
268
269 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
270 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
271
272 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
273 quantifier applied to a forward-referencing subroutine call, could compile
274 incorrect code or give the error "internal error: previously-checked
275 referenced subpattern not found".
276
277 6. Both MS Visual Studio and Symbian OS have problems with initializing
278 variables to point to external functions. For these systems, therefore,
279 pcre_malloc etc. are now initialized to local functions that call the
280 relevant global functions.
281
282 7. There were two entries missing in the vectors called coptable and poptable
283 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
284 I've fixed the data, and added a kludgy way of testing at compile time that
285 the lengths are correct (equal to the number of opcodes).
286
287 8. Following on from 7, I added a similar kludge to check the length of the
288 eint vector in pcreposix.c.
289
290 9. Error texts for pcre_compile() are held as one long string to avoid too
291 much relocation at load time. To find a text, the string is searched,
292 counting zeros. There was no check for running off the end of the string,
293 which could happen if a new error number was added without updating the
294 string.
295
296 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
297
298 11. \K was not working if it appeared in an atomic group or in a group that
299 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
300 \K is "not well defined" if used in an assertion. PCRE now accepts it if
301 the assertion is positive, but not if it is negative.
302
303 12. Change 11 fortuitously reduced the size of the stack frame used in the
304 "match()" function of pcre_exec.c by one pointer. Forthcoming
305 implementation of support for (*MARK) will need an extra pointer on the
306 stack; I have reserved it now, so that the stack frame size does not
307 decrease.
308
309 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
310 item in branch that calls a recursion is a subroutine call - as in the
311 second branch in the above example - was incorrectly given the compile-
312 time error "recursive call could loop indefinitely" because pcre_compile()
313 was not correctly checking the subroutine for matching a non-empty string.
314
315 14. The checks for overrunning compiling workspace could trigger after an
316 overrun had occurred. This is a "should never occur" error, but it can be
317 triggered by pathological patterns such as hundreds of nested parentheses.
318 The checks now trigger 100 bytes before the end of the workspace.
319
320 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
321
322
323 Version 8.01 19-Jan-2010
324 ------------------------
325
326 1. If a pattern contained a conditional subpattern with only one branch (in
327 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
328 computed the wrong minimum data length (which is of course zero for such
329 subpatterns). This could cause incorrect "no match" results.
330
331 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
332 the pattern is reset in the first branch, pcre_compile() failed with
333 "internal error: code overflow at offset...". This happened only when
334 the reset was to the original external option setting. (An optimization
335 abstracts leading options settings into an external setting, which was the
336 cause of this.)
337
338 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
339 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
340 assertion pattern did not match (meaning that the assertion was true), it
341 was incorrectly treated as false if the SKIP had been reached during the
342 matching. This also applied to assertions used as conditions.
343
344 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
345 assertion subpattern, including such a pattern used as a condition,
346 unpredictable results occurred, instead of the error return
347 PCRE_ERROR_DFA_UITEM.
348
349 5. The C++ GlobalReplace function was not working like Perl for the special
350 situation when an empty string is matched. It now does the fancy magic
351 stuff that is necessary.
352
353 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
354 removed. (These were left over from very, very early versions of PCRE.)
355
356 7. Some cosmetic changes to the code to make life easier when compiling it
357 as part of something else:
358
359 (a) Change DEBUG to PCRE_DEBUG.
360
361 (b) In pcre_compile(), rename the member of the "branch_chain" structure
362 called "current" as "current_branch", to prevent a collision with the
363 Linux macro when compiled as a kernel module.
364
365 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
366 prevent a collision with the Linux macro when compiled as a kernel
367 module.
368
369 8. In pcre_compile() there are some checks for integer overflows that used to
370 cast potentially large values to (double). This has been changed to that
371 when building, a check for int64_t is made, and if it is found, it is used
372 instead, thus avoiding the use of floating point arithmetic. (There is no
373 other use of FP in PCRE.) If int64_t is not found, the fallback is to
374 double.
375
376 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
377 2005 (difference between two addresses compared to an unsigned value).
378
379 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
380 custom one, because of the following reported problem in Windows:
381
382 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
383 under Win32.
384 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
385 therefore missing the function definition.
386 - The compiler thus generates a "C" signature for the test function.
387 - The linker fails to find the "C" function.
388 - PCRE fails to configure if asked to do so against libbz2.
389
390 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
391 messages were output:
392
393 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
394 rerunning libtoolize, to keep the correct libtool macros in-tree.
395 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
396
397 I have done both of these things.
398
399 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
400 most of the time, it *can* run out if it is given a pattern that contains a
401 runaway infinite recursion. I updated the discussion in the pcrestack man
402 page.
403
404 13. Now that we have gone to the x.xx style of version numbers, the minor
405 version may start with zero. Using 08 or 09 is a bad idea because users
406 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
407 interpreted as invalid octal numbers. I've updated the previous comment in
408 configure.ac, and also added a check that gives an error if 08 or 09 are
409 used.
410
411 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
412 causing partial matching to fail when the end of the subject matched \W
413 in a UTF-8 pattern where \W was quantified with a minimum of 3.
414
415 15. There were some discrepancies between the declarations in pcre_internal.h
416 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
417 their definitions. The declarations used "const uschar *" and the
418 definitions used USPTR. Even though USPTR is normally defined as "const
419 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
420 reported that: "This difference in casting confuses some C++ compilers, for
421 example, SunCC recognizes above declarations as different functions and
422 generates broken code for hbpcre." I have changed the declarations to use
423 USPTR.
424
425 16. GNU libtool is named differently on some systems. The autogen.sh script now
426 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
427 (FreeBSD).
428
429 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
430 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
431 comment: "Figure out how to create a longlong from a string: strtoll and
432 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
433 instance, but it only takes 2 args instead of 3!"
434
435 18. A subtle bug concerned with back references has been fixed by a change of
436 specification, with a corresponding code fix. A pattern such as
437 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
438 refers, was giving matches when it shouldn't. For example, xa=xaaa would
439 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
440 same bug. Such groups have to be quantified to be useful, or contained
441 inside another quantified group. (If there's no repetition, the reference
442 can never match.) The problem arises because, having left the group and
443 moved on to the rest of the pattern, a later failure that backtracks into
444 the group uses the captured value from the final iteration of the group
445 rather than the correct earlier one. I have fixed this in PCRE by forcing
446 any group that contains a reference to itself to be an atomic group; that
447 is, there cannot be any backtracking into it once it has completed. This is
448 similar to recursive and subroutine calls.
449
450
451 Version 8.00 19-Oct-09
452 ----------------------
453
454 1. The table for translating pcre_compile() error codes into POSIX error codes
455 was out-of-date, and there was no check on the pcre_compile() error code
456 being within the table. This could lead to an OK return being given in
457 error.
458
459 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
460 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
461 in a Windows environment.
462
463 3. The pcregrep --count option prints the count for each file even when it is
464 zero, as does GNU grep. However, pcregrep was also printing all files when
465 --files-with-matches was added. Now, when both options are given, it prints
466 counts only for those files that have at least one match. (GNU grep just
467 prints the file name in this circumstance, but including the count seems
468 more useful - otherwise, why use --count?) Also ensured that the
469 combination -clh just lists non-zero counts, with no names.
470
471 4. The long form of the pcregrep -F option was incorrectly implemented as
472 --fixed_strings instead of --fixed-strings. This is an incompatible change,
473 but it seems right to fix it, and I didn't think it was worth preserving
474 the old behaviour.
475
476 5. The command line items --regex=pattern and --regexp=pattern were not
477 recognized by pcregrep, which required --regex pattern or --regexp pattern
478 (with a space rather than an '='). The man page documented the '=' forms,
479 which are compatible with GNU grep; these now work.
480
481 6. No libpcreposix.pc file was created for pkg-config; there was just
482 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
483
484 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
485 when UCP support is not needed, by modifying the Python script that
486 generates it from Unicode data files. This should not matter if the module
487 is correctly used as a library, but I received one complaint about 50K of
488 unwanted data. My guess is that the person linked everything into his
489 program rather than using a library. Anyway, it does no harm.
490
491 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
492 was a minimum greater than 1 for a wide character in a possessive
493 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
494 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
495 character. Chaos in the form of incorrect output or a compiling loop could
496 result.
497
498 9. The restrictions on what a pattern can contain when partial matching is
499 requested for pcre_exec() have been removed. All patterns can now be
500 partially matched by this function. In addition, if there are at least two
501 slots in the offset vector, the offset of the earliest inspected character
502 for the match and the offset of the end of the subject are set in them when
503 PCRE_ERROR_PARTIAL is returned.
504
505 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
506 synonymous with PCRE_PARTIAL, for backwards compatibility, and
507 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
508 and may be more useful for multi-segment matching.
509
510 11. Partial matching with pcre_exec() is now more intuitive. A partial match
511 used to be given if ever the end of the subject was reached; now it is
512 given only if matching could not proceed because another character was
513 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
514 string "Z", which now yields "no match" instead of "partial match". In the
515 case of pcre_dfa_exec(), "no match" is given if every matching path for the
516 final character ended with (*FAIL).
517
518 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
519 if the pattern had a "must contain" character that was already found in the
520 earlier partial match, unless partial matching was again requested. For
521 example, with the pattern /dog.(body)?/, the "must contain" character is
522 "g". If the first part-match was for the string "dog", restarting with
523 "sbody" failed. This bug has been fixed.
524
525 13. The string returned by pcre_dfa_exec() after a partial match has been
526 changed so that it starts at the first inspected character rather than the
527 first character of the match. This makes a difference only if the pattern
528 starts with a lookbehind assertion or \b or \B (\K is not supported by
529 pcre_dfa_exec()). It's an incompatible change, but it makes the two
530 matching functions compatible, and I think it's the right thing to do.
531
532 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
533 so that the demonstration program is easily available in environments where
534 PCRE has not been installed from source.
535
536 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
537 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
538 library.
539
540 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
541 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
542 is not the first non-POSIX option to be added. Clearly some people find
543 these options useful.
544
545 17. If a caller to the POSIX matching function regexec() passes a non-zero
546 value for nmatch with a NULL value for pmatch, the value of
547 nmatch is forced to zero.
548
549 18. RunGrepTest did not have a test for the availability of the -u option of
550 the diff command, as RunTest does. It now checks in the same way as
551 RunTest, and also checks for the -b option.
552
553 19. If an odd number of negated classes containing just a single character
554 interposed, within parentheses, between a forward reference to a named
555 subpattern and the definition of the subpattern, compilation crashed with
556 an internal error, complaining that it could not find the referenced
557 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
558 [The bug was that it was starting one character too far in when skipping
559 over the character class, thus treating the ] as data rather than
560 terminating the class. This meant it could skip too much.]
561
562 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
563 /g option in pcretest when the pattern contains \K, which makes it possible
564 to have an empty string match not at the start, even when the pattern is
565 anchored. Updated pcretest and pcredemo to use this option.
566
567 21. If the maximum number of capturing subpatterns in a recursion was greater
568 than the maximum at the outer level, the higher number was returned, but
569 with unset values at the outer level. The correct (outer level) value is
570 now given.
571
572 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
573 PCRE did not set those parentheses (unlike Perl). I have now found a way to
574 make it do so. The string so far is captured, making this feature
575 compatible with Perl.
576
577 23. The tests have been re-organized, adding tests 11 and 12, to make it
578 possible to check the Perl 5.10 features against Perl 5.10.
579
580 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
581 pattern matches a fixed length string. PCRE did not allow this; now it
582 does. Neither allows recursion.
583
584 25. I finally figured out how to implement a request to provide the minimum
585 length of subject string that was needed in order to match a given pattern.
586 (It was back references and recursion that I had previously got hung up
587 on.) This code has now been added to pcre_study(); it finds a lower bound
588 to the length of subject needed. It is not necessarily the greatest lower
589 bound, but using it to avoid searching strings that are too short does give
590 some useful speed-ups. The value is available to calling programs via
591 pcre_fullinfo().
592
593 26. While implementing 25, I discovered to my embarrassment that pcretest had
594 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
595 study optimizations had never been tested with that matching function.
596 Oops. What is worse, even when it was passed study data, there was a bug in
597 pcre_dfa_exec() that meant it never actually used it. Double oops. There
598 were also very few tests of studied patterns with pcre_dfa_exec().
599
600 27. If (?| is used to create subpatterns with duplicate numbers, they are now
601 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
602 on the other side of the coin, they are no longer allowed to have different
603 names, because these cannot be distinguished in PCRE, and this has caused
604 confusion. (This is a difference from Perl.)
605
606 28. When duplicate subpattern names are present (necessarily with different
607 numbers, as required by 27 above), and a test is made by name in a
608 conditional pattern, either for a subpattern having been matched, or for
609 recursion in such a pattern, all the associated numbered subpatterns are
610 tested, and the overall condition is true if the condition is true for any
611 one of them. This is the way Perl works, and is also more like the way
612 testing by number works.
613
614
615 Version 7.9 11-Apr-09
616 ---------------------
617
618 1. When building with support for bzlib/zlib (pcregrep) and/or readline
619 (pcretest), all targets were linked against these libraries. This included
620 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
621 libraries. This caused unwanted dependencies to be created. This problem
622 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
623 pcretest is linked with readline.
624
625 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
626 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
627 moved outside it again, because FALSE and TRUE are already defined in AIX,
628 but BOOL is not.
629
630 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
631 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
632
633 4. The pcregrep documentation said spaces were inserted as well as colons (or
634 hyphens) following file names and line numbers when outputting matching
635 lines. This is not true; no spaces are inserted. I have also clarified the
636 wording for the --colour (or --color) option.
637
638 5. In pcregrep, when --colour was used with -o, the list of matching strings
639 was not coloured; this is different to GNU grep, so I have changed it to be
640 the same.
641
642 6. When --colo(u)r was used in pcregrep, only the first matching substring in
643 each matching line was coloured. Now it goes on to look for further matches
644 of any of the test patterns, which is the same behaviour as GNU grep.
645
646 7. A pattern that could match an empty string could cause pcregrep to loop; it
647 doesn't make sense to accept an empty string match in pcregrep, so I have
648 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
649 seems to be how GNU grep behaves.
650
651 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
652 start or after a newline", because the conditional assertion was not being
653 correctly handled. The rule now is that both the assertion and what follows
654 in the first alternative must satisfy the test.
655
656 9. If auto-callout was enabled in a pattern with a conditional group whose
657 condition was an assertion, PCRE could crash during matching, both with
658 pcre_exec() and pcre_dfa_exec().
659
660 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
661 used for matching.
662
663 11. Unicode property support in character classes was not working for
664 characters (bytes) greater than 127 when not in UTF-8 mode.
665
666 12. Added the -M command line option to pcretest.
667
668 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
669
670 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
671
672 16. Added comments and documentation about mis-use of no_arg in the C++
673 wrapper.
674
675 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
676 from Martin Jerabek that uses macro names for all relevant character and
677 string constants.
678
679 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
680 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
681 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
682 these, but not everybody uses configure.
683
684 19. A conditional group that had only one branch was not being correctly
685 recognized as an item that could match an empty string. This meant that an
686 enclosing group might also not be so recognized, causing infinite looping
687 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
688 with the subject "ab", where knowledge that the repeated group can match
689 nothing is needed in order to break the loop.
690
691 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
692 exec(), but without supplying a callout function, matching went wrong.
693
694 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
695 leak if the size of the offset vector was greater than 30. When the vector
696 is smaller, the saved offsets during recursion go onto a local stack
697 vector, but for larger vectors malloc() is used. It was failing to free
698 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
699 error, in fact).
700
701 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
702 heapframe that is used only when UTF-8 support is enabled. This caused no
703 problem, but was untidy.
704
705 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
706 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
707 included within another project.
708
709 24. Steven Van Ingelgem's patches to add more options to the CMake support,
710 slightly modified by me:
711
712 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
713 not building pcregrep.
714
715 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
716 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
717
718 25. Forward references, both numeric and by name, in patterns that made use of
719 duplicate group numbers, could behave incorrectly or give incorrect errors,
720 because when scanning forward to find the reference group, PCRE was not
721 taking into account the duplicate group numbers. A pattern such as
722 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
723
724 26. Changed a few more instances of "const unsigned char *" to USPTR, making
725 the feature of a custom pointer more persuasive (as requested by a user).
726
727 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
728 pcretest.c, inside #ifndefs, because it seems they are sometimes already
729 pre-defined.
730
731 28. Added support for (*UTF8) at the start of a pattern.
732
733 29. Arrange for flags added by the "release type" setting in CMake to be shown
734 in the configuration summary.
735
736
737 Version 7.8 05-Sep-08
738 ---------------------
739
740 1. Replaced UCP searching code with optimized version as implemented for Ad
741 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
742 stage table and inline lookup instead of a function, giving speed ups of 2
743 to 5 times on some simple patterns that I tested. Permission was given to
744 distribute the MultiStage2.py script that generates the tables (it's not in
745 the tarball, but is in the Subversion repository).
746
747 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
748 scripts.
749
750 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
751 a group with a zero qualifier. The result of the study could be incorrect,
752 or the function might crash, depending on the pattern.
753
754 4. Caseless matching was not working for non-ASCII characters in back
755 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
756 It now works when Unicode Property Support is available.
757
758 5. In pcretest, an escape such as \x{de} in the data was always generating
759 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
760 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
761 truncation.
762
763 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
764
765 7. Added two (int) casts to pcregrep when printing the difference of two
766 pointers, in case they are 64-bit values.
767
768 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
769 test 2 if it fails.
770
771 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
772 and a #define of that name to empty if it is not externally set. This is to
773 allow users of MSVC to set it if necessary.
774
775 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
776 the convenience functions in the pcre_get.c source file.
777
778 11. An option change at the start of a pattern that had top-level alternatives
779 could cause overwriting and/or a crash. This command provoked a crash in
780 some environments:
781
782 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
783
784 This potential security problem was recorded as CVE-2008-2371.
785
786 12. For a pattern where the match had to start at the beginning or immediately
787 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
788 pcre_dfa_exec() could read past the end of the passed subject if there was
789 no match. To help with detecting such bugs (e.g. with valgrind), I modified
790 pcretest so that it places the subject at the end of its malloc-ed buffer.
791
792 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
793 exec() might read past the end of the data buffer in UTF-8 mode.
794
795 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
796 the data contained the byte 0x85 as part of a UTF-8 character within its
797 first line. This applied both to normal and DFA matching.
798
799 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
800 /^[^d]*?$/8 failed to match "abc".
801
802 16. Added a missing copyright notice to pcrecpp_internal.h.
803
804 17. Make it more clear in the documentation that values returned from
805 pcre_exec() in ovector are byte offsets, not character counts.
806
807 18. Tidied a few places to stop certain compilers from issuing warnings.
808
809 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
810 supplied by Stefan Weber. I made a further small update for 7.8 because
811 there is a change of source arrangements: the pcre_searchfuncs.c module is
812 replaced by pcre_ucd.c.
813
814
815 Version 7.7 07-May-08
816 ---------------------
817
818 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
819 a string to a long long, pretend we don't even have a long long." This is
820 done by checking for the strtoq, strtoll, and _strtoi64 functions.
821
822 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
823 pre-7.6 versions, which defined a global no_arg variable instead of putting
824 it in the RE class. (See also #8 below.)
825
826 3. Remove a line of dead code, identified by coverity and reported by Nuno
827 Lopes.
828
829 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
830
831 (1) The include/exclude patterns were being applied to the whole pathnames
832 of files, instead of just to the final components.
833
834 (2) If there was more than one level of directory, the subdirectories were
835 skipped unless they satisfied the include/exclude conditions. This is
836 inconsistent with GNU grep (and could even be seen as contrary to the
837 pcregrep specification - which I improved to make it absolutely clear).
838 The action now is always to scan all levels of directory, and just
839 apply the include/exclude patterns to regular files.
840
841 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
842 --exclude_dir in the tests to avoid scanning .svn directories.
843
844 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
845 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
846 doesn't support NULs in patterns.
847
848 7. Added some missing "const"s to declarations of static tables in
849 pcre_compile.c and pcre_dfa_exec.c.
850
851 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
852 caused by fix #2 above. (Subsequently also a second patch to fix the
853 first patch. And a third patch - this was a messy problem.)
854
855 9. Applied Craig's patch to remove the use of push_back().
856
857 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
858 matching function regexec().
859
860 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
861 which, however, unlike Perl's \g{...}, are subroutine calls, not back
862 references. PCRE supports relative numbers with this syntax (I don't think
863 Oniguruma does).
864
865 12. Previously, a group with a zero repeat such as (...){0} was completely
866 omitted from the compiled regex. However, this means that if the group
867 was called as a subroutine from elsewhere in the pattern, things went wrong
868 (an internal error was given). Such groups are now left in the compiled
869 pattern, with a new opcode that causes them to be skipped at execution
870 time.
871
872 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
873 to the way PCRE behaves:
874
875 (a) A lone ] character is dis-allowed (Perl treats it as data).
876
877 (b) A back reference to an unmatched subpattern matches an empty string
878 (Perl fails the current match path).
879
880 (c) A data ] in a character class must be notated as \] because if the
881 first data character in a class is ], it defines an empty class. (In
882 Perl it is not possible to have an empty class.) The empty class []
883 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
884 The negative empty class [^] matches any one character, independently
885 of the DOTALL setting.
886
887 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
888 non-existent subpattern following a character class starting with ']' and
889 containing () gave an internal compiling error instead of "reference to
890 non-existent subpattern". Fortunately, when the pattern did exist, the
891 compiled code was correct. (When scanning forwards to check for the
892 existencd of the subpattern, it was treating the data ']' as terminating
893 the class, so got the count wrong. When actually compiling, the reference
894 was subsequently set up correctly.)
895
896 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
897 it was being rejected as not supported by pcre_dfa_exec(), even though
898 other assertions are supported. I have made pcre_dfa_exec() support
899 (*FAIL).
900
901 16. The implementation of 13c above involved the invention of a new opcode,
902 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
903 cannot be changed at match time, I realized I could make a small
904 improvement to matching performance by compiling OP_ALLANY instead of
905 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
906 on the OP_ANY path.
907
908 17. Compiling pcretest on Windows with readline support failed without the
909 following two fixes: (1) Make the unistd.h include conditional on
910 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
911
912 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
913 ncurses library to be included for pcretest when ReadLine support is
914 requested, but also to allow for it to be overridden. This patch came from
915 Daniel Bergström.
916
917 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
918 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
919 any errors with the current Unicode tables. Thanks to Peter Kankowski for
920 spotting this.
921
922
923 Version 7.6 28-Jan-08
924 ---------------------
925
926 1. A character class containing a very large number of characters with
927 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
928 overflow.
929
930 2. Patch to cut out the "long long" test in pcrecpp_unittest when
931 HAVE_LONG_LONG is not defined.
932
933 3. Applied Christian Ehrlicher's patch to update the CMake build files to
934 bring them up to date and include new features. This patch includes:
935
936 - Fixed PH's badly added libz and libbz2 support.
937 - Fixed a problem with static linking.
938 - Added pcredemo. [But later removed - see 7 below.]
939 - Fixed dftables problem and added an option.
940 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
941 HAVE_LONG_LONG.
942 - Added readline support for pcretest.
943 - Added an listing of the option settings after cmake has run.
944
945 4. A user submitted a patch to Makefile that makes it easy to create
946 "pcre.dll" under mingw when using Configure/Make. I added stuff to
947 Makefile.am that cause it to include this special target, without
948 affecting anything else. Note that the same mingw target plus all
949 the other distribution libraries and programs are now supported
950 when configuring with CMake (see 6 below) instead of with
951 Configure/Make.
952
953 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
954 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
955 exported in the Windows port". It has not yet been confirmed that the patch
956 solves the problem, but it does no harm.
957
958 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
959 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
960 with CMake, and also correct the comment about stack recursion.
961
962 7. Remove the automatic building of pcredemo from the ./configure system and
963 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
964 of a program that users should build themselves after PCRE is installed, so
965 building it automatically is not really right. What is more, it gave
966 trouble in some build environments.
967
968 8. Further tidies to CMakeLists.txt from Sheri and Christian.
969
970
971 Version 7.5 10-Jan-08
972 ---------------------
973
974 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
975 values in parens when parsing an RE using the C++ wrapper."
976
977 2. Negative specials like \S did not work in character classes in UTF-8 mode.
978 Characters greater than 255 were excluded from the class instead of being
979 included.
980
981 3. The same bug as (2) above applied to negated POSIX classes such as
982 [:^space:].
983
984 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
985 defined or documented. It seems to have been a typo for PCRE_STATIC, so
986 I have changed it.
987
988 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
989 first named subpattern) and a construct such as (?&a) would reference the
990 first named subpattern whose name started with "a" (in other words, the
991 length check was missing). Both these problems are fixed. "Subpattern name
992 expected" is now given for (?&) (a zero-length name), and this patch also
993 makes it give the same error for \k'' (previously it complained that that
994 was a reference to a non-existent subpattern).
995
996 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
997 this is right because (?- can be followed by option settings as well as by
998 digits. I have, however, made the messages clearer.
999
1000 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1001 than the number used in the conditional) now cause a compile-time error.
1002 This is actually not compatible with Perl, which accepts such patterns, but
1003 treats the conditional as always being FALSE (as PCRE used to), but it
1004 seems to me that giving a diagnostic is better.
1005
1006 8. Change "alphameric" to the more common word "alphanumeric" in comments
1007 and messages.
1008
1009 9. Fix two occurrences of "backslash" in comments that should have been
1010 "backspace".
1011
1012 10. Remove two redundant lines of code that can never be obeyed (their function
1013 was moved elsewhere).
1014
1015 11. The program that makes PCRE's Unicode character property table had a bug
1016 which caused it to generate incorrect table entries for sequences of
1017 characters that have the same character type, but are in different scripts.
1018 It amalgamated them into a single range, with the script of the first of
1019 them. In other words, some characters were in the wrong script. There were
1020 thirteen such cases, affecting characters in the following ranges:
1021
1022 U+002b0 - U+002c1
1023 U+0060c - U+0060d
1024 U+0061e - U+00612
1025 U+0064b - U+0065e
1026 U+0074d - U+0076d
1027 U+01800 - U+01805
1028 U+01d00 - U+01d77
1029 U+01d9b - U+01dbf
1030 U+0200b - U+0200f
1031 U+030fc - U+030fe
1032 U+03260 - U+0327f
1033 U+0fb46 - U+0fbb1
1034 U+10450 - U+1049d
1035
1036 12. The -o option (show only the matching part of a line) for pcregrep was not
1037 compatible with GNU grep in that, if there was more than one match in a
1038 line, it showed only the first of them. It now behaves in the same way as
1039 GNU grep.
1040
1041 13. If the -o and -v options were combined for pcregrep, it printed a blank
1042 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1043 does the same. The return code can be used to tell if there were any
1044 non-matching lines.
1045
1046 14. Added --file-offsets and --line-offsets to pcregrep.
1047
1048 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1049 infinitely looping recursion. The bug was that positive lookaheads were not
1050 being skipped when checking for a possible empty match (negative lookaheads
1051 and both kinds of lookbehind were skipped).
1052
1053 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1054 inclusion of <windows.h> to before rather than after the definition of
1055 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1056
1057 17. Specifying a possessive quantifier with a specific limit for a Unicode
1058 character property caused pcre_compile() to compile bad code, which led at
1059 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1060 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1061 caused the error; without that there was no problem.
1062
1063 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1064
1065 19. Added --enable-pcretest-libreadline.
1066
1067 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1068 RE::GlobalReplace(). As a result, the number of replacements returned was
1069 double what it should be. I removed one of the increments, but Craig sent a
1070 later patch that removed the other one (the right fix) and added unit tests
1071 that check the return values (which was not done before).
1072
1073 21. Several CMake things:
1074
1075 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1076 the names libpcre and libpcreposix, not just pcre and pcreposix.
1077
1078 (2) The above change means that pcretest and pcregrep are now correctly
1079 linked with the newly-built libraries, not previously installed ones.
1080
1081 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1082
1083 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1084 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1085 UTF-8 newline character). The key issue is that the pattern starts .*;
1086 this means that the match must be either at the beginning, or after a
1087 newline. The bug was in the code for advancing after a failed match and
1088 checking that the new position followed a newline. It was not taking
1089 account of UTF-8 characters correctly.
1090
1091 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1092 character classes. PCRE was not treating the sequence [:...:] as a
1093 character class unless the ... were all letters. Perl, however, seems to
1094 allow any characters between [: and :], though of course it rejects as
1095 unknown any "names" that contain non-letters, because all the known class
1096 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1097 for example, whereas PCRE did not - it did not recognize a POSIX character
1098 class. This seemed a bit dangerous, so the code has been changed to be
1099 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1100 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1101 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1102 Perl does, and where it didn't before.
1103
1104 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1105 Windows environments %n is disabled by default.
1106
1107
1108 Version 7.4 21-Sep-07
1109 ---------------------
1110
1111 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1112 means that a class such as [\s] counted as "explicit reference to CR or
1113 LF". That isn't really right - the whole point of the change was to try to
1114 help when there was an actual mention of one of the two characters. So now
1115 the change happens only if \r or \n (or a literal CR or LF) character is
1116 encountered.
1117
1118 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1119 of both had grown to the point where there were only 3 bits left.
1120 Fortunately, there was spare space in the data structure, and so I have
1121 moved the internal flags into a new 16-bit field to free up more option
1122 bits.
1123
1124 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1125 but did not set the internal JCHANGED flag - either of these is enough to
1126 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1127 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1128 start sets both bits.
1129
1130 4. Added options (at build time, compile time, exec time) to change \R from
1131 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1132
1133 5. doc/pcresyntax.html was missing from the distribution.
1134
1135 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1136 compatibility, even though it is no longer used.
1137
1138 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1139 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1140 windows.h file is present (where different names are used). [This was
1141 reversed later after testing - see 16 below.]
1142
1143 8. Changed all #include <config.h> to #include "config.h". There were also
1144 some further <pcre.h> cases that I changed to "pcre.h".
1145
1146 9. When pcregrep was used with the --colour option, it missed the line ending
1147 sequence off the lines that it output.
1148
1149 10. It was pointed out to me that arrays of string pointers cause lots of
1150 relocations when a shared library is dynamically loaded. A technique of
1151 using a single long string with a table of offsets can drastically reduce
1152 these. I have refactored PCRE in four places to do this. The result is
1153 dramatic:
1154
1155 Originally: 290
1156 After changing UCP table: 187
1157 After changing error message table: 43
1158 After changing table of "verbs" 36
1159 After changing table of Posix names 22
1160
1161 Thanks to the folks working on Gregex for glib for this insight.
1162
1163 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1164 unicode-properties was also set.
1165
1166 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1167
1168 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1169 checked only for CRLF.
1170
1171 14. Added casts to pcretest.c to avoid compiler warnings.
1172
1173 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1174
1175 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1176 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1177 entirely. This removes changes made in 7 above.
1178
1179 17. The CMake files have been updated, and there is now more information about
1180 building with CMake in the NON-UNIX-USE document.
1181
1182
1183 Version 7.3 28-Aug-07
1184 ---------------------
1185
1186 1. In the rejigging of the build system that eventually resulted in 7.1, the
1187 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1188 brackets there is not right, since it causes compilers to look for an
1189 installed pcre.h, not the version that is in the source that is being
1190 compiled (which of course may be different). I have changed it back to:
1191
1192 #include "pcre.h"
1193
1194 I have a vague recollection that the change was concerned with compiling in
1195 different directories, but in the new build system, that is taken care of
1196 by the VPATH setting the Makefile.
1197
1198 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1199 when the subject happened to end in the byte 0x85 (e.g. if the last
1200 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1201 characters but of course it shouldn't be taken as a newline when it is part
1202 of another character. The bug was that, for an unlimited repeat of . in
1203 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1204 characters when looking for a newline.
1205
1206 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1207
1208 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1209 in debug output.
1210
1211 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1212 long printing in the pcrecpp unittest when running under MinGW.
1213
1214 6. ESC_K was left out of the EBCDIC table.
1215
1216 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1217 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1218 limit also applies to "virtual nesting" when a pattern is recursive, and in
1219 this case 1000 isn't so big. I have been able to remove this limit at the
1220 expense of backing off one optimization in certain circumstances. Normally,
1221 when pcre_exec() would call its internal match() function recursively and
1222 immediately return the result unconditionally, it uses a "tail recursion"
1223 feature to save stack. However, when a subpattern that can match an empty
1224 string has an unlimited repetition quantifier, it no longer makes this
1225 optimization. That gives it a stack frame in which to save the data for
1226 checking that an empty string has been matched. Previously this was taken
1227 from the 1000-entry workspace that had been reserved. So now there is no
1228 explicit limit, but more stack is used.
1229
1230 8. Applied Daniel's patches to solve problems with the import/export magic
1231 syntax that is required for Windows, and which was going wrong for the
1232 pcreposix and pcrecpp parts of the library. These were overlooked when this
1233 problem was solved for the main library.
1234
1235 9. There were some crude static tests to avoid integer overflow when computing
1236 the size of patterns that contain repeated groups with explicit upper
1237 limits. As the maximum quantifier is 65535, the maximum group length was
1238 set at 30,000 so that the product of these two numbers did not overflow a
1239 32-bit integer. However, it turns out that people want to use groups that
1240 are longer than 30,000 bytes (though not repeat them that many times).
1241 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1242 made it possible to implement the integer overflow checks in a much more
1243 dynamic way, which I have now done. The artificial limitation on group
1244 length has been removed - we now have only the limit on the total length of
1245 the compiled pattern, which depends on the LINK_SIZE setting.
1246
1247 10. Fixed a bug in the documentation for get/copy named substring when
1248 duplicate names are permitted. If none of the named substrings are set, the
1249 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1250 empty string.
1251
1252 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1253 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1254 because the ] is interpreted as the first data character and the
1255 terminating ] is not found. PCRE has been made compatible with Perl in this
1256 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1257 cause memory overwriting.
1258
1259 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1260 string has been matched (to stop an infinite loop). It was not recognizing
1261 a conditional subpattern that could match an empty string if that
1262 subpattern was within another subpattern. For example, it looped when
1263 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1264 condition was not nested. This bug has been fixed.
1265
1266 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1267 past the start of the subject in the presence of bytes with the top bit
1268 set, for example "\x8aBCD".
1269
1270 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1271 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1272
1273 14. Optimized (?!) to (*FAIL).
1274
1275 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1276 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1277 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1278 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1279 does: it's just the validity check that is more restrictive.
1280
1281 16. Inserted checks for integer overflows during escape sequence (backslash)
1282 processing, and also fixed erroneous offset values for syntax errors during
1283 backslash processing.
1284
1285 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1286 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1287
1288 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1289 caused an overrun.
1290
1291 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1292 something other than just ASCII characters) inside a group that had an
1293 unlimited repeat caused a loop at compile time (while checking to see
1294 whether the group could match an empty string).
1295
1296 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1297 [\P{Any}] did so. (Error in the code for printing property names.)
1298
1299 21. An orphan \E inside a character class could cause a crash.
1300
1301 22. A repeated capturing bracket such as (A)? could cause a wild memory
1302 reference during compilation.
1303
1304 23. There are several functions in pcre_compile() that scan along a compiled
1305 expression for various reasons (e.g. to see if it's fixed length for look
1306 behind). There were bugs in these functions when a repeated \p or \P was
1307 present in the pattern. These operators have additional parameters compared
1308 with \d, etc, and these were not being taken into account when moving along
1309 the compiled data. Specifically:
1310
1311 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1312 length.
1313
1314 (b) An item such as \pL+ within a repeated group could cause crashes or
1315 loops.
1316
1317 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1318 "reference to non-existent subpattern" error.
1319
1320 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1321
1322 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1323 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1324
1325 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1326
1327 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1328 character were causing crashes (broken optimization).
1329
1330 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1331 \p or \P) caused a compile-time loop.
1332
1333 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1334 break. For example, the unstudied pattern [\r\n]A does not match the string
1335 "\r\nA" because change 7.0/46 below moves the current point on by two
1336 characters after failing to match at the start. However, the pattern \nA
1337 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1338 the same is true. There doesn't seem any very clean way out of this, but
1339 what I have chosen to do makes the common cases work: PCRE now takes note
1340 of whether there can be an explicit match for \r or \n anywhere in the
1341 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1342 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1343 pattern has explicit CR or LF references.
1344
1345 29. Added (*CR) etc for changing newline setting at start of pattern.
1346
1347
1348 Version 7.2 19-Jun-07
1349 ---------------------
1350
1351 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1352 which is apparently normally available under Windows.
1353
1354 2. Re-jig the pcregrep tests with different newline settings in an attempt
1355 to make them independent of the local environment's newline setting.
1356
1357 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1358
1359 4. Some of the "internals" tests were previously cut out when the link size
1360 was not 2, because the output contained actual offsets. The recent new
1361 "Z" feature of pcretest means that these can be cut out, making the tests
1362 usable with all link sizes.
1363
1364 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1365 stack recursion. This gives a massive performance boost under BSD, but just
1366 a small improvement under Linux. However, it saves one field in the frame
1367 in all cases.
1368
1369 6. Added more features from the forthcoming Perl 5.10:
1370
1371 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1372 recursion call. It refers to the nth most recently opened parentheses.
1373
1374 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1375 to be opened parentheses.
1376
1377 (c) Conditions that refer to capturing parentheses can be specified
1378 relatively, for example, (?(-2)... or (?(+3)...
1379
1380 (d) \K resets the start of the current match so that everything before
1381 is not part of it.
1382
1383 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1384
1385 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1386 reference syntax.
1387
1388 (g) (?| introduces a group in which the numbering of parentheses in each
1389 alternative starts with the same number.
1390
1391 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1392
1393 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1394 PCRE_INFO_JCHANGED.
1395
1396 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1397 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1398 for detecting groups that can match an empty string.
1399
1400 9. A pattern with a very large number of alternatives (more than several
1401 hundred) was running out of internal workspace during the pre-compile
1402 phase, where pcre_compile() figures out how much memory will be needed. A
1403 bit of new cunning has reduced the workspace needed for groups with
1404 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1405 workspace instead of running out of the 4096 that are available.
1406
1407 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1408
1409 11. Applied patch from Google to remove an optimization that didn't quite work.
1410 The report of the bug said:
1411
1412 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1413 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1414 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1415
1416 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1417 it matched the wrong number of bytes.
1418
1419
1420 Version 7.1 24-Apr-07
1421 ---------------------
1422
1423 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1424 that is more "standard", making use of automake and other Autotools. There
1425 is some re-arrangement of the files and adjustment of comments consequent
1426 on this.
1427
1428 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1429 for recursive directory scanning broke on some systems because the files
1430 are not scanned in any specific order and on different systems the order
1431 was different. A call to "sort" has been inserted into RunGrepTest for the
1432 approprate test as a short-term fix. In the longer term there may be an
1433 alternative.
1434
1435 3. I had an email from Eric Raymond about problems translating some of PCRE's
1436 man pages to HTML (despite the fact that I distribute HTML pages, some
1437 people do their own conversions for various reasons). The problems
1438 concerned the use of low-level troff macros .br and .in. I have therefore
1439 removed all such uses from the man pages (some were redundant, some could
1440 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1441 HTML has been updated to handle .nf/.fi and to complain if it encounters
1442 .br or .in.
1443
1444 4. Updated comments in configure.ac that get placed in config.h.in and also
1445 arranged for config.h to be included in the distribution, with the name
1446 config.h.generic, for the benefit of those who have to compile without
1447 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1448
1449 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1450 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1451 makevp.bat for the current PCRE, using the additional files
1452 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1453
1454 6. A Windows user reported a minor discrepancy with test 2, which turned out
1455 to be caused by a trailing space on an input line that had got lost in his
1456 copy. The trailing space was an accident, so I've just removed it.
1457
1458 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1459 that is needed.
1460
1461 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1462 as "const" (a) because they are and (b) because it helps the PHP
1463 maintainers who have recently made a script to detect big data structures
1464 in the php code that should be moved to the .rodata section. I remembered
1465 to update Builducptable as well, so it won't revert if ucptable.h is ever
1466 re-created.
1467
1468 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1469 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1470 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1471 support is not required. This saves 1.5-2K of code, which is important in
1472 some applications.
1473
1474 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1475 so as not to refer to the tables, even though these functions will never be
1476 called when UTF-8 support is disabled. Otherwise there are problems with a
1477 shared library.
1478
1479 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1480
1481 (a) It was defining its arguments as char * instead of void *.
1482
1483 (b) It was assuming that all moves were upwards in memory; this was true
1484 a long time ago when I wrote it, but is no longer the case.
1485
1486 The emulated memove() is provided for those environments that have neither
1487 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1488 is clearly not the case, as these two bugs were recently reported.
1489
1490 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1491 and Detrail to create the HTML documentation, the .txt form of the man
1492 pages, and it removes trailing spaces from listed files. It also creates
1493 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1494 case, it wraps all the #defines with #ifndefs. This script should be run
1495 before "make dist".
1496
1497 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1498 with Unicode property support.
1499
1500 (a) For a maximizing quantifier, if the two different cases of the
1501 character were of different lengths in their UTF-8 codings (there are
1502 some cases like this - I found 11), and the matching function had to
1503 back up over a mixture of the two cases, it incorrectly assumed they
1504 were both the same length.
1505
1506 (b) When PCRE was configured to use the heap rather than the stack for
1507 recursion during matching, it was not correctly preserving the data for
1508 the other case of a UTF-8 character when checking ahead for a match
1509 while processing a minimizing repeat. If the check also involved
1510 matching a wide character, but failed, corruption could cause an
1511 erroneous result when trying to check for a repeat of the original
1512 character.
1513
1514 13. Some tidying changes to the testing mechanism:
1515
1516 (a) The RunTest script now detects the internal link size and whether there
1517 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1518 values substituted by "configure". (The RunGrepTest script already did
1519 this for UTF-8.) The configure.ac script no longer substitutes the
1520 relevant variables.
1521
1522 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1523 with length and offset values. This means that the output is different
1524 for different internal link sizes. Test 2 is skipped for link sizes
1525 other than 2 because of this, bypassing the problem. Unfortunately,
1526 there was also a test in test 3 (the locale tests) that used /B and
1527 failed for link sizes other than 2. Rather than cut the whole test out,
1528 I have added a new /Z option to pcretest that replaces the length and
1529 offset values with spaces. This is now used to make test 3 independent
1530 of link size. (Test 2 will be tidied up later.)
1531
1532 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1533 segmentation fault instead of returning the appropriate error message.
1534
1535 15. In multiline mode when the newline sequence was set to "any", the pattern
1536 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1537 This doesn't seem right; it now treats the CRLF combination as the line
1538 ending, and so does not match in that case. It's only a pattern such as ^$
1539 that would hit this one: something like ^ABC$ would have failed after \r
1540 and then tried again after \r\n.
1541
1542 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1543 in an attempt to make files that differ only in their line terminators
1544 compare equal. This works on Linux.
1545
1546 17. Under certain error circumstances pcregrep might try to free random memory
1547 as it exited. This is now fixed, thanks to valgrind.
1548
1549 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1550 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1551 was because its rules for how to advance for /g after matching an empty
1552 string at the end of a line did not allow for this case. They now check for
1553 it specially.
1554
1555 20. pcretest is supposed to handle patterns and data of any length, by
1556 extending its buffers when necessary. It was getting this wrong when the
1557 buffer for a data line had to be extended.
1558
1559 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1560 CRLF as a newline sequence.
1561
1562 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1563 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1564 I have nevertheless tidied it up.
1565
1566 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1567
1568 24. Added a man page for pcre-config.
1569
1570
1571 Version 7.0 19-Dec-06
1572 ---------------------
1573
1574 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1575 moving to gcc 4.1.1.
1576
1577 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1578 sys/time.h, which is documented as needed for this function. It doesn't
1579 seem to matter on Linux, but it showed up on some releases of OS X.
1580
1581 3. It seems that there are systems where bytes whose values are greater than
1582 127 match isprint() in the "C" locale. The "C" locale should be the
1583 default when a C program starts up. In most systems, only ASCII printing
1584 characters match isprint(). This difference caused the output from pcretest
1585 to vary, making some of the tests fail. I have changed pcretest so that:
1586
1587 (a) When it is outputting text in the compiled version of a pattern, bytes
1588 other than 32-126 are always shown as hex escapes.
1589
1590 (b) When it is outputting text that is a matched part of a subject string,
1591 it does the same, unless a different locale has been set for the match
1592 (using the /L modifier). In this case, it uses isprint() to decide.
1593
1594 4. Fixed a major bug that caused incorrect computation of the amount of memory
1595 required for a compiled pattern when options that changed within the
1596 pattern affected the logic of the preliminary scan that determines the
1597 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1598 that the computed length was too small. The symptoms of this bug were
1599 either the PCRE error "internal error: code overflow" from pcre_compile(),
1600 or a glibc crash with a message such as "pcretest: free(): invalid next
1601 size (fast)". Examples of patterns that provoked this bug (shown in
1602 pcretest format) are:
1603
1604 /(?-x: )/x
1605 /(?x)(?-x: \s*#\s*)/
1606 /((?i)[\x{c0}])/8
1607 /(?i:[\x{c0}])/8
1608
1609 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1610 is now done differently.
1611
1612 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1613 wrapper classes; (b) implement a new function in the C++ scanner that is
1614 more efficient than the old way of doing things because it avoids levels of
1615 recursion in the regex matching; (c) add a paragraph to the documentation
1616 for the FullMatch() function.
1617
1618 6. The escape sequence \n was being treated as whatever was defined as
1619 "newline". Not only was this contrary to the documentation, which states
1620 that \n is character 10 (hex 0A), but it also went horribly wrong when
1621 "newline" was defined as CRLF. This has been fixed.
1622
1623 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1624 was being set to -1 for the "end of line" case (supposedly a value that no
1625 character can have). Though this value is never used (the check for end of
1626 line is "zero bytes in current character"), it caused compiler complaints.
1627 I've changed it to 0xffffffff.
1628
1629 8. In pcre_version.c, the version string was being built by a sequence of
1630 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1631 string (as it is for production releases) called a macro with an empty
1632 argument. The C standard says the result of this is undefined. The gcc
1633 compiler treats it as an empty string (which was what was wanted) but it is
1634 reported that Visual C gives an error. The source has been hacked around to
1635 avoid this problem.
1636
1637 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1638 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1639 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1640 of them did).
1641
1642 10. Originally, pcretest opened its input and output without "b"; then I was
1643 told that "b" was needed in some environments, so it was added for release
1644 5.0 to both the input and output. (It makes no difference on Unix-like
1645 systems.) Later I was told that it is wrong for the input on Windows. I've
1646 now abstracted the modes into two macros, to make it easier to fiddle with
1647 them, and removed "b" from the input mode under Windows.
1648
1649 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1650
1651 12. Added -help and --help to pcretest as an official way of being reminded
1652 of the options.
1653
1654 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1655 and pcrecpp.cc because they annoy compilers at high warning levels.
1656
1657 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1658
1659 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1660 scripts are not C programs :-) and which was not noticed because it works
1661 on Linux.
1662
1663 16. pcretest is supposed to handle any length of pattern and data line (as one
1664 line or as a continued sequence of lines) by extending its input buffer if
1665 necessary. This feature was broken for very long pattern lines, leading to
1666 a string of junk being passed to pcre_compile() if the pattern was longer
1667 than about 50K.
1668
1669 17. I have done a major re-factoring of the way pcre_compile() computes the
1670 amount of memory needed for a compiled pattern. Previously, there was code
1671 that made a preliminary scan of the pattern in order to do this. That was
1672 OK when PCRE was new, but as the facilities have expanded, it has become
1673 harder and harder to keep it in step with the real compile phase, and there
1674 have been a number of bugs (see for example, 4 above). I have now found a
1675 cunning way of running the real compile function in a "fake" mode that
1676 enables it to compute how much memory it would need, while actually only
1677 ever using a few hundred bytes of working memory and without too many
1678 tests of the mode. This should make future maintenance and development
1679 easier. A side effect of this work is that the limit of 200 on the nesting
1680 depth of parentheses has been removed (though this was never a serious
1681 limitation, I suspect). However, there is a downside: pcre_compile() now
1682 runs more slowly than before (30% or more, depending on the pattern). I
1683 hope this isn't a big issue. There is no effect on runtime performance.
1684
1685 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1686 newline (only possible for the last line of a file) and it was a
1687 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1688
1689 19. Added additional timing features to pcretest. (1) The -tm option now times
1690 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1691 separate command line item, by a number that specifies the number of
1692 repeats to use when timing. The default is 50000; this gives better
1693 precision, but takes uncomfortably long for very large patterns.
1694
1695 20. Extended pcre_study() to be more clever in cases where a branch of a
1696 subpattern has no definite first character. For example, (a*|b*)[cd] would
1697 previously give no result from pcre_study(). Now it recognizes that the
1698 first character must be a, b, c, or d.
1699
1700 21. There was an incorrect error "recursive call could loop indefinitely" if
1701 a subpattern (or the entire pattern) that was being tested for matching an
1702 empty string contained only one non-empty item after a nested subpattern.
1703 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1704 incorrectly, because the \d was being skipped in the check.
1705
1706 22. The pcretest program now has a new pattern option /B and a command line
1707 option -b, which is equivalent to adding /B to every pattern. This causes
1708 it to show the compiled bytecode, without the additional information that
1709 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1710 is the same as /B/I).
1711
1712 23. A new optimization is now able automatically to treat some sequences such
1713 as a*b as a*+b. More specifically, if something simple (such as a character
1714 or a simple class like \d) has an unlimited quantifier, and is followed by
1715 something that cannot possibly match the quantified thing, the quantifier
1716 is automatically "possessified".
1717
1718 24. A recursive reference to a subpattern whose number was greater than 39
1719 went wrong under certain circumstances in UTF-8 mode. This bug could also
1720 have affected the operation of pcre_study().
1721
1722 25. Realized that a little bit of performance could be had by replacing
1723 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1724
1725 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1726
1727 27. Possessive quantifiers such as a++ were previously implemented by turning
1728 them into atomic groups such as ($>a+). Now they have their own opcodes,
1729 which improves performance. This includes the automatically created ones
1730 from 23 above.
1731
1732 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1733 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1734 the first matched character to be a colon. This applied both to named and
1735 numbered groups.
1736
1737 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1738
1739 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1740 building PCRE on Windows easier, so I have included it in the distribution.
1741
1742 31. There is now a check in pcretest against a ridiculously large number being
1743 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1744 loop, the loop is abandoned.
1745
1746 32. Forward references to subpatterns in conditions such as (?(2)...) where
1747 subpattern 2 is defined later cause pcre_compile() to search forwards in
1748 the pattern for the relevant set of parentheses. This search went wrong
1749 when there were unescaped parentheses in a character class, parentheses
1750 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1751
1752 33. "Subroutine" calls and backreferences were previously restricted to
1753 referencing subpatterns earlier in the regex. This restriction has now
1754 been removed.
1755
1756 34. Added a number of extra features that are going to be in Perl 5.10. On the
1757 whole, these are just syntactic alternatives for features that PCRE had
1758 previously implemented using the Python syntax or my own invention. The
1759 other formats are all retained for compatibility.
1760
1761 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1762 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1763 also .NET compatible.
1764
1765 (b) A recursion or subroutine call to a named group can now be defined as
1766 (?&name) as well as (?P>name).
1767
1768 (c) A backreference to a named group can now be defined as \k<name> or
1769 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1770 5.10, are also .NET compatible.
1771
1772 (d) A conditional reference to a named group can now use the syntax
1773 (?(<name>) or (?('name') as well as (?(name).
1774
1775 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1776 groups (named and numbered) that are never evaluated inline, but can be
1777 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1778 is always false. There may be only one alternative in such a group.
1779
1780 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1781 as the simple (?(R). The condition is true only if the most recent
1782 recursion is that of the given number or name. It does not search out
1783 through the entire recursion stack.
1784
1785 (g) The escape \gN or \g{N} has been added, where N is a positive or
1786 negative number, specifying an absolute or relative reference.
1787
1788 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1789 some "unreachable code" warnings.
1790
1791 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1792 things, this adds five new scripts.
1793
1794 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1795 There were also incompatibilities regarding the handling of \Q..\E inside
1796 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1797 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1798
1799 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1800 matches an empty string, and forcibly breaks the loop. There were bugs in
1801 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1802 against aaaa the result was just "a" rather than "aaaa", for example. Two
1803 separate and independent bugs (that affected different cases) have been
1804 fixed.
1805
1806 39. Refactored the code to abolish the use of different opcodes for small
1807 capturing bracket numbers. This is a tidy that I avoided doing when I
1808 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1809 The new approach is not only tidier, it makes it possible to reduce the
1810 memory needed to fix the previous bug (38).
1811
1812 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1813 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1814 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1815 mode.
1816
1817 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1818 report.
1819
1820 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1821 copy construction and assignment in the C++ wrapper.
1822
1823 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1824 couple of bugs that could have given wrong results in the "--newline=crlf"
1825 case.
1826
1827 44. Added a number of casts and did some reorganization of signed/unsigned int
1828 variables following suggestions from Dair Grant. Also renamed the variable
1829 "this" as "item" because it is a C++ keyword.
1830
1831 45. Arranged for dftables to add
1832
1833 #include "pcre_internal.h"
1834
1835 to pcre_chartables.c because without it, gcc 4.x may remove the array
1836 definition from the final binary if PCRE is built into a static library and
1837 dead code stripping is activated.
1838
1839 46. For an unanchored pattern, if a match attempt fails at the start of a
1840 newline sequence, and the newline setting is CRLF or ANY, and the next two
1841 characters are CRLF, advance by two characters instead of one.
1842
1843
1844 Version 6.7 04-Jul-06
1845 ---------------------
1846
1847 1. In order to handle tests when input lines are enormously long, pcretest has
1848 been re-factored so that it automatically extends its buffers when
1849 necessary. The code is crude, but this _is_ just a test program. The
1850 default size has been increased from 32K to 50K.
1851
1852 2. The code in pcre_study() was using the value of the re argument before
1853 testing it for NULL. (Of course, in any sensible call of the function, it
1854 won't be NULL.)
1855
1856 3. The memmove() emulation function in pcre_internal.h, which is used on
1857 systems that lack both memmove() and bcopy() - that is, hardly ever -
1858 was missing a "static" storage class specifier.
1859
1860 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1861 containing an extended class (one that cannot be represented by a bitmap
1862 because it contains high-valued characters or Unicode property items, e.g.
1863 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1864 pattern, but PCRE should not loop if you do not (it no longer does).
1865 [Detail: two cases were found: (a) a repeated subpattern containing an
1866 extended class; (b) a recursive reference to a subpattern that followed a
1867 previous extended class. It wasn't skipping over the extended class
1868 correctly when UTF-8 mode was not set.]
1869
1870 5. A negated single-character class was not being recognized as fixed-length
1871 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1872 compile error "lookbehind assertion is not fixed length".
1873
1874 6. The RunPerlTest auxiliary script was showing an unexpected difference
1875 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1876 write a Perl script that can interpret lines of an input file either as
1877 byte characters or as UTF-8, which is what "perltest" was being required to
1878 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1879 can't do is switch easily at run time between having the "use utf8;" pragma
1880 or not. In the end, I fudged it by using the RunPerlTest script to insert
1881 "use utf8;" explicitly for the UTF-8 tests.
1882
1883 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1884 the end of the subject string, contrary to the documentation and to what
1885 Perl does. This was true of both matching functions. Now it matches only at
1886 the start of the subject and immediately after *internal* newlines.
1887
1888 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1889 a pointer to an int instead of a pointer to an unsigned long int. This
1890 caused problems on 64-bit systems.
1891
1892 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1893 instance of the 'standard' template library not being so standard".
1894
1895 10. There was no check on the number of named subpatterns nor the maximum
1896 length of a subpattern name. The product of these values is used to compute
1897 the size of the memory block for a compiled pattern. By supplying a very
1898 long subpattern name and a large number of named subpatterns, the size
1899 computation could be caused to overflow. This is now prevented by limiting
1900 the length of names to 32 characters, and the number of named subpatterns
1901 to 10,000.
1902
1903 11. Subpatterns that are repeated with specific counts have to be replicated in
1904 the compiled pattern. The size of memory for this was computed from the
1905 length of the subpattern and the repeat count. The latter is limited to
1906 65535, but there was no limit on the former, meaning that integer overflow
1907 could in principle occur. The compiled length of a repeated subpattern is
1908 now limited to 30,000 bytes in order to prevent this.
1909
1910 12. Added the optional facility to have named substrings with the same name.
1911
1912 13. Added the ability to use a named substring as a condition, using the
1913 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1914 are numbers (not recommended). Forward references are permitted.
1915
1916 14. Added forward references in named backreferences (if you see what I mean).
1917
1918 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1919 pattern could run off the end of the subject. For example, the pattern
1920 "(?s)(.{1,5})"8 did this with the subject "ab".
1921
1922 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1923 PCRE_CASELESS was set when matching characters that were quantified with ?
1924 or *.
1925
1926 17. A character class other than a single negated character that had a minimum
1927 but no maximum quantifier - for example [ab]{6,} - was not handled
1928 correctly by pce_dfa_exec(). It would match only one character.
1929
1930 18. A valid (though odd) pattern that looked like a POSIX character
1931 class but used an invalid character after [ (for example [[,abc,]]) caused
1932 pcre_compile() to give the error "Failed: internal error: code overflow" or
1933 in some cases to crash with a glibc free() error. This could even happen if
1934 the pattern terminated after [[ but there just happened to be a sequence of
1935 letters, a binary zero, and a closing ] in the memory that followed.
1936
1937 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1938 over the years. Originally (before any Unicode support), just the bottom 8
1939 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1940 output from "man perlunicode" includes this:
1941
1942 The regular expression compiler produces polymorphic opcodes. That
1943 is, the pattern adapts to the data and automatically switches to
1944 the Unicode character scheme when presented with Unicode data--or
1945 instead uses a traditional byte scheme when presented with byte
1946 data.
1947
1948 Sadly, a wide octal escape does not cause a switch, and in a string with
1949 no other multibyte characters, these octal escapes are treated as before.
1950 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1951 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1952 Unicode string.
1953
1954 I have not perpetrated such confusion in PCRE. Up till now, it took just
1955 the bottom 8 bits, as in old Perl. I have now made octal escapes with
1956 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1957 translate to the appropriate multibyte character.
1958
1959 29. Applied some refactoring to reduce the number of warnings from Microsoft
1960 and Borland compilers. This has included removing the fudge introduced
1961 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1962 a warning about an unused variable.
1963
1964 21. PCRE has not included VT (character 0x0b) in the set of whitespace
1965 characters since release 4.0, because Perl (from release 5.004) does not.
1966 [Or at least, is documented not to: some releases seem to be in conflict
1967 with the documentation.] However, when a pattern was studied with
1968 pcre_study() and all its branches started with \s, PCRE still included VT
1969 as a possible starting character. Of course, this did no harm; it just
1970 caused an unnecessary match attempt.
1971
1972 22. Removed a now-redundant internal flag bit that recorded the fact that case
1973 dependency changed within the pattern. This was once needed for "required
1974 byte" processing, but is no longer used. This recovers a now-scarce options
1975 bit. Also moved the least significant internal flag bit to the most-
1976 significant bit of the word, which was not previously used (hangover from
1977 the days when it was an int rather than a uint) to free up another bit for
1978 the future.
1979
1980 23. Added support for CRLF line endings as well as CR and LF. As well as the
1981 default being selectable at build time, it can now be changed at runtime
1982 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1983 specify that it is scanning data with non-default line endings.
1984
1985 24. Changed the definition of CXXLINK to make it agree with the definition of
1986 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1987
1988 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1989 recursions. This makes a big different to stack usage for some patterns.
1990
1991 26. If a subpattern containing a named recursion or subroutine reference such
1992 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1993 the space required for the compiled pattern went wrong and gave too small a
1994 value. Depending on the environment, this could lead to "Failed: internal
1995 error: code overflow at offset 49" or "glibc detected double free or
1996 corruption" errors.
1997
1998 27. Applied patches from Google (a) to support the new newline modes and (b) to
1999 advance over multibyte UTF-8 characters in GlobalReplace.
2000
2001 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2002 difference for some implementation of PCRE in some Windows version.
2003
2004 29. Added some extra testing facilities to pcretest:
2005
2006 \q<number> in a data line sets the "match limit" value
2007 \Q<number> in a data line sets the "match recursion limt" value
2008 -S <number> sets the stack size, where <number> is in megabytes
2009
2010 The -S option isn't available for Windows.
2011
2012
2013 Version 6.6 06-Feb-06
2014 ---------------------
2015
2016 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2017 in pcreposix.h. I have copied the definition from pcre.h.
2018
2019 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2020 because pcre.h is no longer a built file.
2021
2022 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2023 not normally included in the compiled code.
2024
2025
2026 Version 6.5 01-Feb-06
2027 ---------------------
2028
2029 1. When using the partial match feature with pcre_dfa_exec(), it was not
2030 anchoring the second and subsequent partial matches at the new starting
2031 point. This could lead to incorrect results. For example, with the pattern
2032 /1234/, partially matching against "123" and then "a4" gave a match.
2033
2034 2. Changes to pcregrep:
2035
2036 (a) All non-match returns from pcre_exec() were being treated as failures
2037 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2038 error message is output. Some extra information is given for the
2039 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2040 probably the only errors that are likely to be caused by users (by
2041 specifying a regex that has nested indefinite repeats, for instance).
2042 If there are more than 20 of these errors, pcregrep is abandoned.
2043
2044 (b) A binary zero was treated as data while matching, but terminated the
2045 output line if it was written out. This has been fixed: binary zeroes
2046 are now no different to any other data bytes.
2047
2048 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2049 used to set a locale for matching. The --locale=xxxx long option has
2050 been added (no short equivalent) to specify a locale explicitly on the
2051 pcregrep command, overriding the environment variables.
2052
2053 (d) When -B was used with -n, some line numbers in the output were one less
2054 than they should have been.
2055
2056 (e) Added the -o (--only-matching) option.
2057
2058 (f) If -A or -C was used with -c (count only), some lines of context were
2059 accidentally printed for the final match.
2060
2061 (g) Added the -H (--with-filename) option.
2062
2063 (h) The combination of options -rh failed to suppress file names for files
2064 that were found from directory arguments.
2065
2066 (i) Added the -D (--devices) and -d (--directories) options.
2067
2068 (j) Added the -F (--fixed-strings) option.
2069
2070 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2071
2072 (l) Added the --colo(u)r option.
2073
2074 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2075 is not present by default.
2076
2077 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2078 items such as (?R) or (?1), when the recursion could match a number of
2079 alternatives. If it matched one of the alternatives, but subsequently,
2080 outside the recursion, there was a failure, the code tried to back up into
2081 the recursion. However, because of the way PCRE is implemented, this is not
2082 possible, and the result was an incorrect result from the match.
2083
2084 In order to prevent this happening, the specification of recursion has
2085 been changed so that all such subpatterns are automatically treated as
2086 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2087
2088 4. I had overlooked the fact that, in some locales, there are characters for
2089 which isalpha() is true but neither isupper() nor islower() are true. In
2090 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2091 and ordfeminine) are like this. This affected the treatment of \w and \W
2092 when they appeared in character classes, but not when they appeared outside
2093 a character class. The bit map for "word" characters is now created
2094 separately from the results of isalnum() instead of just taking it from the
2095 upper, lower, and digit maps. (Plus the underscore character, of course.)
2096
2097 5. The above bug also affected the handling of POSIX character classes such as
2098 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2099 permanent tables. Instead, the bit maps for such a class were previously
2100 created as the appropriate unions of the upper, lower, and digit bitmaps.
2101 Now they are created by subtraction from the [[:word:]] class, which has
2102 its own bitmap.
2103
2104 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2105 It is created by subtracting the vertical space characters (\x09, \x0a,
2106 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2107 subtraction was done in the overall bitmap for a character class, meaning
2108 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2109 be recognized. This bug has been fixed.
2110
2111 7. Patches from the folks at Google:
2112
2113 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2114 real life, but is still worth protecting against".
2115
2116 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2117 regular expressions".
2118
2119 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2120 have it.
2121
2122 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2123 "configure" and the latter not, in order to fix a problem somebody had
2124 with compiling the Arg class on HP-UX.
2125
2126 (e) Improve the error-handling of the C++ wrapper a little bit.
2127
2128 (f) New tests for checking recursion limiting.
2129
2130 8. The pcre_memmove() function, which is used only if the environment does not
2131 have a standard memmove() function (and is therefore rarely compiled),
2132 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2133 returning a result (though PCRE never actually uses the result).
2134
2135 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2136 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2137 returned instead of calling malloc() with an overflowing number that would
2138 most likely cause subsequent chaos.
2139
2140 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2141
2142 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2143 with this option is matched, the nmatch and pmatch options of regexec() are
2144 ignored.
2145
2146 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2147 provided in case anyone wants to the the POSIX interface with UTF-8
2148 strings.
2149
2150 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2151 C++ linking (needed for some HP-UX environments).
2152
2153 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2154 (unused parameter) and in the pcre_printint() function (omitted "default"
2155 switch label when the default is to do nothing).
2156
2157 15. Added some code to make it possible, when PCRE is compiled as a C++
2158 library, to replace subject pointers for pcre_exec() with a smart pointer
2159 class, thus making it possible to process discontinuous strings.
2160
2161 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2162 much the same function. They were added by different people who were trying
2163 to make PCRE easy to compile on non-Unix systems. It has been suggested
2164 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2165 for compiling on Windows systems. I have therefore replaced it with
2166 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2167 defaults to "extern" for C or "extern C" for C++, which works fine on
2168 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2169 SCOPE with something explicit in config.h. In addition:
2170
2171 (a) pcreposix.h still had just "extern" instead of either of these macros;
2172 I have replaced it with PCRE_DATA_SCOPE.
2173
2174 (b) Functions such as _pcre_xclass(), which are internal to the library,
2175 but external in the C sense, all had PCRE_EXPORT in their definitions.
2176 This is apparently wrong for the Windows case, so I have removed it.
2177 (It makes no difference on Unix-like systems.)
2178
2179 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2180 of recursive calls to match(). This is different to MATCH_LIMIT because
2181 that limits the total number of calls to match(), not all of which increase
2182 the depth of recursion. Limiting the recursion depth limits the amount of
2183 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2184 when PCRE is compiled, and changed at run time. A patch from Google adds
2185 this functionality to the C++ interface.
2186
2187 18. Changes to the handling of Unicode character properties:
2188
2189 (a) Updated the table to Unicode 4.1.0.
2190
2191 (b) Recognize characters that are not in the table as "Cn" (undefined).
2192
2193 (c) I revised the way the table is implemented to a much improved format
2194 which includes recognition of ranges. It now supports the ranges that
2195 are defined in UnicodeData.txt, and it also amalgamates other
2196 characters into ranges. This has reduced the number of entries in the
2197 table from around 16,000 to around 3,000, thus reducing its size
2198 considerably. I realized I did not need to use a tree structure after
2199 all - a binary chop search is just as efficient. Having reduced the
2200 number of entries, I extended their size from 6 bytes to 8 bytes to
2201 allow for more data.
2202
2203 (d) Added support for Unicode script names via properties such as \p{Han}.
2204
2205 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2206 matching that character.
2207
2208 20. When matching a repeated Unicode property with a minimum greater than zero,
2209 (for example \pL{2,}), PCRE could look past the end of the subject if it
2210 reached it while seeking the minimum number of characters. This could
2211 happen only if some of the characters were more than one byte long, because
2212 there is a check for at least the minimum number of bytes.
2213
2214 21. Refactored the implementation of \p and \P so as to be more general, to
2215 allow for more different types of property in future. This has changed the
2216 compiled form incompatibly. Anybody with saved compiled patterns that use
2217 \p or \P will have to recompile them.
2218
2219 22. Added "Any" and "L&" to the supported property types.
2220
2221 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2222 but give a compile time error if the value is greater than 0xff.
2223
2224 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2225 accidentally not being installed or uninstalled.
2226
2227 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2228 made were to insert the current release number. This seemed silly, because
2229 it made things harder for people building PCRE on systems that don't run
2230 "configure". I have turned pcre.h into a distributed file, no longer built
2231 by "configure", with the version identification directly included. There is
2232 no longer a pcre.h.in file.
2233
2234 However, this change necessitated a change to the pcre-config script as
2235 well. It is built from pcre-config.in, and one of the substitutions was the
2236 release number. I have updated configure.ac so that ./configure now finds
2237 the release number by grepping pcre.h.
2238
2239 26. Added the ability to run the tests under valgrind.
2240
2241
2242 Version 6.4 05-Sep-05
2243 ---------------------
2244
2245 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2246 "--" to be printed when multiple files were scanned, even when none of the
2247 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2248 consider it to be a bug, and have restored the previous behaviour.
2249
2250 2. A couple of code tidies to get rid of compiler warnings.
2251
2252 3. The pcretest program used to cheat by referring to symbols in the library
2253 whose names begin with _pcre_. These are internal symbols that are not
2254 really supposed to be visible externally, and in some environments it is
2255 possible to suppress them. The cheating is now confined to including
2256 certain files from the library's source, which is a bit cleaner.
2257
2258 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2259 file's purpose clearer.
2260
2261 5. Reorganized pcre_ucp_findchar().
2262
2263
2264 Version 6.3 15-Aug-05
2265 ---------------------
2266
2267 1. The file libpcre.pc.in did not have general read permission in the tarball.
2268
2269 2. There were some problems when building without C++ support:
2270
2271 (a) If C++ support was not built, "make install" and "make test" still
2272 tried to test it.
2273
2274 (b) There were problems when the value of CXX was explicitly set. Some
2275 changes have been made to try to fix these, and ...
2276
2277 (c) --disable-cpp can now be used to explicitly disable C++ support.
2278
2279 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2280 backslash in a target when C++ was disabled. This confuses some
2281 versions of "make", apparently. Using an intermediate variable solves
2282 this. (Same for CPP_LOBJ.)
2283
2284 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2285 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2286 necessary on certain architectures.
2287
2288 4. Added a setting of -export-symbols-regex to the link command to remove
2289 those symbols that are exported in the C sense, but actually are local
2290 within the library, and not documented. Their names all begin with
2291 "_pcre_". This is not a perfect job, because (a) we have to except some
2292 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2293 available (and never for static libraries). I have made a note to try to
2294 find a way round (a) in the future.
2295
2296
2297 Version 6.2 01-Aug-05
2298 ---------------------
2299
2300 1. There was no test for integer overflow of quantifier values. A construction
2301 such as {1111111111111111} would give undefined results. What is worse, if
2302 a minimum quantifier for a parenthesized subpattern overflowed and became
2303 negative, the calculation of the memory size went wrong. This could have
2304 led to memory overwriting.
2305
2306 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2307
2308 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2309 operating environments where this matters.
2310
2311 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2312 PCRE options from within the C++ wrapper.
2313
2314 5. Named capturing subpatterns were not being correctly counted when a pattern
2315 was compiled. This caused two problems: (a) If there were more than 100
2316 such subpatterns, the calculation of the memory needed for the whole
2317 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2318 back references of the form \12, where the number was greater than 9, were
2319 not recognized as back references, even though there were sufficient
2320 previous subpatterns.
2321
2322 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2323 versions of gcc, e.g. 2.95.4.
2324
2325
2326 Version 6.1 21-Jun-05
2327 ---------------------
2328
2329 1. There was one reference to the variable "posix" in pcretest.c that was not
2330 surrounded by "#if !defined NOPOSIX".
2331
2332 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2333 the cross-check on the old pcre_info() function, for the benefit of the
2334 cut-down version of PCRE that is currently imported into Exim.
2335
2336 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2337 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2338 patterns that start (?i) but I don't think that matters. The use of (?i) is
2339 just an example; this all applies to the other options as well.
2340
2341 4. Since libtool seems to echo the compile commands it is issuing, the output
2342 from "make" can be reduced a bit by putting "@" in front of each libtool
2343 compile command.
2344
2345 5. Patch from the folks at Google for configure.in to be a bit more thorough
2346 in checking for a suitable C++ installation before trying to compile the
2347 C++ stuff. This should fix a reported problem when a compiler was present,
2348 but no suitable headers.
2349
2350 6. The man pages all had just "PCRE" as their title. I have changed them to
2351 be the relevant file name. I have also arranged that these names are
2352 retained in the file doc/pcre.txt, which is a concatenation in text format
2353 of all the man pages except the little individual ones for each function.
2354
2355 7. The NON-UNIX-USE file had not been updated for the different set of source
2356 files that come with release 6. I also added a few comments about the C++
2357 wrapper.
2358
2359
2360 Version 6.0 07-Jun-05
2361 ---------------------
2362
2363 1. Some minor internal re-organization to help with my DFA experiments.
2364
2365 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2366 didn't matter for the library itself when fully configured, but did matter
2367 when compiling without UCP support, or within Exim, where the ucp files are
2368 not imported.
2369
2370 3. Refactoring of the library code to split up the various functions into
2371 different source modules. The addition of the new DFA matching code (see
2372 below) to a single monolithic source would have made it really too
2373 unwieldy, quite apart from causing all the code to be include in a
2374 statically linked application, when only some functions are used. This is
2375 relevant even without the DFA addition now that patterns can be compiled in
2376 one application and matched in another.
2377
2378 The downside of splitting up is that there have to be some external
2379 functions and data tables that are used internally in different modules of
2380 the library but which are not part of the API. These have all had their
2381 names changed to start with "_pcre_" so that they are unlikely to clash
2382 with other external names.
2383
2384 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2385 a different (DFA) algorithm. Although it is slower than the original
2386 function, it does have some advantages for certain types of matching
2387 problem.
2388
2389 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2390 including restarting after a partial match.
2391
2392 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2393 defined when compiling for Windows was sent to me. I have put it into the
2394 code, though I have no means of testing or verifying it.
2395
2396 7. Added the pcre_refcount() auxiliary function.
2397
2398 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2399 match before or at the first newline in the subject string. In pcretest,
2400 the /f option on a pattern can be used to set this.
2401
2402 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2403 would behave wrongly. This has been present in PCRE since release 4.0.
2404
2405 10. A number of changes to the pcregrep command:
2406
2407 (a) Refactored how -x works; insert ^(...)$ instead of setting
2408 PCRE_ANCHORED and checking the length, in preparation for adding
2409 something similar for -w.
2410
2411 (b) Added the -w (match as a word) option.
2412
2413 (c) Refactored the way lines are read and buffered so as to have more
2414 than one at a time available.
2415
2416 (d) Implemented a pcregrep test script.
2417
2418 (e) Added the -M (multiline match) option. This allows patterns to match
2419 over several lines of the subject. The buffering ensures that at least
2420 8K, or the rest of the document (whichever is the shorter) is available
2421 for matching (and similarly the previous 8K for lookbehind assertions).
2422
2423 (f) Changed the --help output so that it now says
2424
2425 -w, --word-regex(p)
2426
2427 instead of two lines, one with "regex" and the other with "regexp"
2428 because that confused at least one person since the short forms are the
2429 same. (This required a bit of code, as the output is generated
2430 automatically from a table. It wasn't just a text change.)
2431
2432 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2433 option but starts with a hyphen. Could be a pattern or a path name
2434 starting with a hyphen, for instance.
2435
2436 (h) "-" can be given as a file name to represent stdin.
2437
2438 (i) When file names are being printed, "(standard input)" is used for
2439 the standard input, for compatibility with GNU grep. Previously
2440 "<stdin>" was used.
2441
2442 (j) The option --label=xxx can be used to supply a name to be used for
2443 stdin when file names are being printed. There is no short form.
2444
2445 (k) Re-factored the options decoding logic because we are going to add
2446 two more options that take data. Such options can now be given in four
2447 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2448
2449 (l) Added the -A, -B, and -C options for requesting that lines of context
2450 around matches be printed.
2451
2452 (m) Added the -L option to print the names of files that do not contain
2453 any matching lines, that is, the complement of -l.
2454
2455 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2456 continue to scan other files.
2457
2458 (o) The -s option was incorrectly implemented. For compatibility with other
2459 greps, it now suppresses the error message for a non-existent or non-
2460 accessible file (but not the return code). There is a new option called
2461 -q that suppresses the output of matching lines, which was what -s was
2462 previously doing.
2463
2464 (p) Added --include and --exclude options to specify files for inclusion
2465 and exclusion when recursing.
2466
2467 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2468 Hopefully, it now does.
2469
2470 12. Missing cast in pcre_study().
2471
2472 13. Added an "uninstall" target to the makefile.
2473
2474 14. Replaced "extern" in the function prototypes in Makefile.in with
2475 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2476 world, but is set differently for Windows.
2477
2478 15. Added a second compiling function called pcre_compile2(). The only
2479 difference is that it has an extra argument, which is a pointer to an
2480 integer error code. When there is a compile-time failure, this is set
2481 non-zero, in addition to the error test pointer being set to point to an
2482 error message. The new argument may be NULL if no error number is required
2483 (but then you may as well call pcre_compile(), which is now just a
2484 wrapper). This facility is provided because some applications need a
2485 numeric error indication, but it has also enabled me to tidy up the way
2486 compile-time errors are handled in the POSIX wrapper.
2487
2488 16. Added VPATH=.libs to the makefile; this should help when building with one
2489 prefix path and installing with another. (Or so I'm told by someone who
2490 knows more about this stuff than I do.)
2491
2492 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2493 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2494 match everything, including newlines. This is not POSIX-compatible, but
2495 somebody wanted the feature. From pcretest it can be activated by using
2496 both the P and the s flags.
2497
2498 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2499
2500 19. libpcre.pc was being incorrectly installed as executable.
2501
2502 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2503 it now also looks for '\r' so that it will work unmodified on Windows.
2504
2505 21. Added Google's contributed C++ wrapper to the distribution.
2506
2507 22. Added some untidy missing memory free() calls in pcretest, to keep
2508 Electric Fence happy when testing.
2509
2510
2511
2512 Version 5.0 13-Sep-04
2513 ---------------------
2514
2515 1. Internal change: literal characters are no longer packed up into items
2516 containing multiple characters in a single byte-string. Each character
2517 is now matched using a separate opcode. However, there may be more than one
2518 byte in the character in UTF-8 mode.
2519
2520 2. The pcre_callout_block structure has two new fields: pattern_position and
2521 next_item_length. These contain the offset in the pattern to the next match
2522 item, and its length, respectively.
2523
2524 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2525 insertion of callouts before each pattern item. Added the /C option to
2526 pcretest to make use of this.
2527
2528 4. On the advice of a Windows user, the lines
2529
2530 #if defined(_WIN32) || defined(WIN32)
2531 _setmode( _fileno( stdout ), 0x8000 );
2532 #endif /* defined(_WIN32) || defined(WIN32) */
2533
2534 have been added to the source of pcretest. This apparently does useful
2535 magic in relation to line terminators.
2536
2537 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2538 for the benefit of those environments where the "b" makes a difference.
2539
2540 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2541 to know about it. I have put a hack into configure.in that adds in code
2542 to set GCC=yes if CC=icc. This seems to end up at a point in the
2543 generated configure script that is early enough to affect the setting of
2544 compiler options, which is what is needed, but I have no means of testing
2545 whether it really works. (The user who reported this had patched the
2546 generated configure script, which of course I cannot do.)
2547
2548 LATER: After change 22 below (new libtool files), the configure script
2549 seems to know about icc (and also ecc). Therefore, I have commented out
2550 this hack in configure.in.
2551
2552 7. Added support for pkg-config (2 patches were sent in).
2553
2554 8. Negated POSIX character classes that used a combination of internal tables
2555 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2556 [[:^ascii]]. Typically, they would match almost any characters. The other
2557 POSIX classes were not broken in this way.
2558
2559 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2560 to find the match, as PCRE was deluded into thinking that the match had to
2561 start at the start point or following a newline. The same bug applied to
2562 patterns with negative forward assertions or any backward assertions
2563 preceding ".*" at the start, unless the pattern required a fixed first
2564 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2565
2566 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2567 starting at the last subject character, bytes beyond the end of the subject
2568 string were read.
2569
2570 11. Renamed the variable "class" as "classbits" to make life easier for C++
2571 users. (Previously there was a macro definition, but it apparently wasn't
2572 enough.)
2573
2574 12. Added the new field "tables" to the extra data so that tables can be passed
2575 in at exec time, or the internal tables can be re-selected. This allows
2576 a compiled regex to be saved and re-used at a later time by a different
2577 program that might have everything at different addresses.
2578
2579 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2580 -R library as well as a -L library.
2581
2582 14. The debugging options of pcretest (-d on the command line or D on a
2583 pattern) showed incorrect output for anything following an extended class
2584 that contained multibyte characters and which was followed by a quantifier.
2585
2586 15. Added optional support for general category Unicode character properties
2587 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2588 support. It adds about 90K to the size of the library. The meanings of the
2589 inbuilt class escapes such as \d and \s have NOT been changed.
2590
2591 16. Updated pcredemo.c to include calls to free() to release the memory for the
2592 compiled pattern.
2593
2594 17. The generated file chartables.c was being created in the source directory
2595 instead of in the building directory. This caused the build to fail if the
2596 source directory was different from the building directory, and was
2597 read-only.
2598
2599 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2600 file. No doubt somebody will tell me if they don't make sense... Also added
2601 Dan Mooney's comments about building on OpenVMS.
2602
2603 19. Added support for partial matching via the PCRE_PARTIAL option for
2604 pcre_exec() and the \P data escape in pcretest.
2605
2606 20. Extended pcretest with 3 new pattern features:
2607
2608 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2609 write the compiled pattern to the file whose name is "rest-of-line".
2610 This is a straight binary dump of the data, with the saved pointer to
2611 the character tables forced to be NULL. The study data, if any, is
2612 written too. After writing, pcretest reads a new pattern.
2613
2614 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2615 compiled pattern from the given file. There must not be any
2616 occurrences of "<" in the file name (pretty unlikely); if there are,
2617 pcretest will instead treat the initial "<" as a pattern delimiter.
2618 After reading in the pattern, pcretest goes on to read data lines as
2619 usual.
2620
2621 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2622 and 16-bit fields in a compiled pattern, to simulate a pattern that
2623 was compiled on a host of opposite endianness.
2624
2625 21. The pcre-exec() function can now cope with patterns that were compiled on
2626 hosts of opposite endianness, with this restriction:
2627
2628 As for any compiled expression that is saved and used later, the tables
2629 pointer field cannot be preserved; the extra_data field in the arguments
2630 to pcre_exec() should be used to pass in a tables address if a value
2631 other than the default internal tables were used at compile time.
2632
2633 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2634 now diagnosed as an error. Previously, most of the time, a negative number
2635 would have been treated as zero, but if in addition "ovector" was passed as
2636 NULL, a crash could occur.
2637
2638 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2639 new versions from the libtool 1.5 distribution (the last one is a copy of
2640 a file called libtool.m4). This seems to have fixed the need to patch
2641 "configure" to support Darwin 1.3 (which I used to do). However, I still
2642 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2643 workstation).
2644
2645 24. Changed the PCRE licence to be the more standard "BSD" licence.
2646
2647
2648 Version 4.5 01-Dec-03
2649 ---------------------
2650
2651 1. There has been some re-arrangement of the code for the match() function so
2652 that it can be compiled in a version that does not call itself recursively.
2653 Instead, it keeps those local variables that need separate instances for
2654 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2655 needs to "recurse". Keeping track of where control must go is done by means
2656 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2657 hide most of the details from the main code, and operates only if
2658 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2659 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2660 operating.
2661
2662 To make it easier for callers to provide specially tailored get/free
2663 functions for this usage, two new functions, pcre_stack_malloc, and
2664 pcre_stack_free, are used. They are always called in strict stacking order,
2665 and the size of block requested is always the same.
2666
2667 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2668 PCRE has been compiled to use the stack or the heap for recursion. The
2669 -C option of pcretest uses this to show which version is compiled.
2670
2671 A new data escape \S, is added to pcretest; it causes the amounts of store
2672 obtained and freed by both kinds of malloc/free at match time to be added
2673 to the output.
2674
2675 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2676 what's available on my current Linux desktop machine.
2677
2678 3. When matching a UTF-8 string, the test for a valid string at the start has
2679 been extended. If start_offset is not zero, PCRE now checks that it points
2680 to a byte that is the start of a UTF-8 character. If not, it returns
2681 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2682 this is necessary because there may be backward assertions in the pattern.
2683 When matching the same subject several times, it may save resources to use
2684 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2685
2686 4. The code for checking the validity of UTF-8 strings has been tightened so
2687 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2688 containing "overlong sequences".
2689
2690 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2691 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2692 should have been "&", but it just so happened that all the cases this let
2693 through by mistake were picked up later in the function.
2694
2695 6. I had used a variable called "isblank" - this is a C99 function, causing
2696 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2697
2698 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2699 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2700 so the version is shown; (c) stop "make test" echoing "./RunTest".
2701
2702 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2703
2704 9. The prototype for memmove() for systems that don't have it was using
2705 size_t, but the inclusion of the header that defines size_t was later. I've
2706 moved the #includes for the C headers earlier to avoid this.
2707
2708 10. Added some adjustments to the code to make it easier to compiler on certain
2709 special systems:
2710
2711 (a) Some "const" qualifiers were missing.
2712 (b) Added the macro EXPORT before all exported functions; by default this
2713 is defined to be empty.
2714 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2715 that it reads its output file name as an argument instead of writing
2716 to the standard output and assuming this can be redirected.
2717
2718 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2719 class containing characters with values greater than 255, PCRE compilation
2720 went into a loop.
2721
2722 12. A recursive reference to a subpattern that was within another subpattern
2723 that had a minimum quantifier of zero caused PCRE to crash. For example,
2724 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2725 recursion. If the recursively-called subpattern itself had a zero repeat,
2726 that was OK.
2727
2728 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2729 buffer into which it was copied (for escape processing) was still set at
2730 1024, so long lines caused crashes.
2731
2732 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2733 "internal error: code overflow...". This applied to any character class
2734 that was followed by a possessive quantifier.
2735
2736 15. Modified the Makefile to add libpcre.la as a prerequisite for
2737 libpcreposix.la because I was told this is needed for a parallel build to
2738 work.
2739
2740 16. If a pattern that contained .* following optional items at the start was
2741 studied, the wrong optimizing data was generated, leading to matching
2742 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2743 matching string must start with a or b or c. The correct conclusion for
2744 this pattern is that a match can start with any character.
2745
2746
2747 Version 4.4 13-Aug-03
2748 ---------------------
2749
2750 1. In UTF-8 mode, a character class containing characters with values between
2751 127 and 255 was not handled correctly if the compiled pattern was studied.
2752 In fixing this, I have also improved the studying algorithm for such
2753 classes (slightly).
2754
2755 2. Three internal functions had redundant arguments passed to them. Removal
2756 might give a very teeny performance improvement.
2757
2758 3. Documentation bug: the value of the capture_top field in a callout is *one
2759 more than* the number of the hightest numbered captured substring.
2760
2761 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2762 in incorrectly linking with a previously installed version. They now link
2763 explicitly with libpcre.la.
2764
2765 5. configure.in no longer needs to recognize Cygwin specially.
2766
2767 6. A problem in pcre.in for Windows platforms is fixed.
2768
2769 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2770 pcretest, it used to include the size of the study block as part of its
2771 output. Unfortunately, the structure contains a field that has a different
2772 size on different hardware architectures. This meant that the tests that
2773 showed this size failed. As the block is currently always of a fixed size,
2774 this information isn't actually particularly useful in pcretest output, so
2775 I have just removed it.
2776
2777 8. Three pre-processor statements accidentally did not start in column 1.
2778 Sadly, there are *still* compilers around that complain, even though
2779 standard C has not required this for well over a decade. Sigh.
2780
2781 9. In pcretest, the code for checking callouts passed small integers in the
2782 callout_data field, which is a void * field. However, some picky compilers
2783 complained about the casts involved for this on 64-bit systems. Now
2784 pcretest passes the address of the small integer instead, which should get
2785 rid of the warnings.
2786
2787 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2788 both compile and run time, and gives an error if an invalid UTF-8 sequence
2789 is found. There is a option for disabling this check in cases where the
2790 string is known to be correct and/or the maximum performance is wanted.
2791
2792 11. In response to a bug report, I changed one line in Makefile.in from
2793
2794 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2795 to
2796 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2797
2798 to look similar to other lines, but I have no way of telling whether this
2799 is the right thing to do, as I do not use Windows. No doubt I'll get told
2800 if it's wrong...
2801
2802
2803 Version 4.3 21-May-03
2804 ---------------------
2805
2806 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2807 Makefile.
2808
2809 2. Some refactoring to improve the quality of the code:
2810
2811 (i) The utf8_table... variables are now declared "const".
2812
2813 (ii) The code for \cx, which used the "case flipping" table to upper case
2814 lower case letters, now just substracts 32. This is ASCII-specific,
2815 but the whole concept of \cx is ASCII-specific, so it seems
2816 reasonable.
2817
2818 (iii) PCRE was using its character types table to recognize decimal and
2819 hexadecimal digits in the pattern. This is silly, because it handles
2820 only 0-9, a-f, and A-F, but the character types table is locale-
2821 specific, which means strange things might happen. A private
2822 table is now used for this - though it costs 256 bytes, a table is
2823 much faster than multiple explicit tests. Of course, the standard
2824 character types table is still used for matching digits in subject
2825 strings against \d.
2826
2827 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2828 ending in _t are). So I've renamed it as ESC_tee.
2829
2830 3. The first argument for regexec() in the POSIX wrapper should have been
2831 defined as "const".
2832
2833 4. Changed pcretest to use malloc() for its buffers so that they can be
2834 Electric Fenced for debugging.
2835
2836 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2837 to read one or more bytes before the start of the subject string. Often this
2838 had no effect on PCRE's behaviour, but in some circumstances it could
2839 provoke a segmentation fault.
2840
2841 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2842 to try to read one or more bytes before the start of the subject string.
2843
2844 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2845 UTF-8 support could misbehave in various ways if the subject string
2846 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2847 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2848 back over UTF-8 characters.)
2849
2850
2851 Version 4.2 14-Apr-03
2852 ---------------------
2853
2854 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2855
2856 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2857 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2858 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2859 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2860 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2861 and BUILD_EXEEXT
2862 Note: automatic setting of the BUILD variables is not yet working
2863 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2864 compile-time but not at link-time
2865 [LINK]: use for linking executables only
2866 make different versions for Windows and non-Windows
2867 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2868 libraries
2869 [LINK_FOR_BUILD]: new variable
2870 [OBJEXT]: use throughout
2871 [EXEEXT]: use throughout
2872 <winshared>: new target
2873 <wininstall>: new target
2874 <dftables.o>: use native compiler
2875 <dftables>: use native linker
2876 <install>: handle Windows platform correctly
2877 <clean>: ditto
2878 <check>: ditto
2879 copy DLL to top builddir before testing
2880
2881 As part of these changes, -no-undefined was removed again. This was reported
2882 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2883 in any case.
2884
2885 3. Some tidies to get rid of compiler warnings:
2886
2887 . In the match_data structure, match_limit was an unsigned long int, whereas
2888 match_call_count was an int. I've made them both unsigned long ints.
2889
2890 . In pcretest the fact that a const uschar * doesn't automatically cast to
2891 a void * provoked a warning.
2892
2893 . Turning on some more compiler warnings threw up some "shadow" variables
2894 and a few more missing casts.
2895
2896 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2897 option, a class that contained a single character with a value between 128
2898 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2899
2900 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2901 option, a class that contained several characters, but with at least one
2902 whose value was between 128 and 255 caused PCRE to crash.
2903
2904
2905 Version 4.1 12-Mar-03
2906 ---------------------
2907
2908 1. Compiling with gcc -pedantic found a couple of places where casts were
2909 needed, and a string in dftables.c that was longer than standard compilers are
2910 required to support.
2911
2912 2. Compiling with Sun's compiler found a few more places where the code could
2913 be tidied up in order to avoid warnings.
2914
2915 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2916 first of these names is deprecated in the latest Autoconf in favour of the name
2917 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2918 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2919 analogy I have changed it to CFLAGS_FOR_BUILD.
2920
2921 4. Added -no-undefined to the linking command in the Makefile, because this is
2922 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2923 linking step for the pcreposix library.
2924
2925 5. PCRE was failing to diagnose the case of two named groups with the same
2926 name.
2927
2928 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2929 literal character that is needed in the subject for a match, and scans along to
2930 ensure that it is present before embarking on the full matching process. This
2931 saves time in cases of nested unlimited repeats that are never going to match.
2932 Problem: the scan can take a lot of time if the subject is very long (e.g.
2933 megabytes), thus penalizing straightforward matches. It is now done only if the
2934 amount of subject to be scanned is less than 1000 bytes.
2935
2936 7. A lesser problem with the same optimization is that it was recording the
2937 first character of an anchored pattern as "needed", thus provoking a search
2938 right along the subject, even when the first match of the pattern was going to
2939 fail. The "needed" character is now not set for anchored patterns, unless it
2940 follows something in the pattern that is of non-fixed length. Thus, it still
2941 fulfils its original purpose of finding quick non-matches in cases of nested
2942 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2943
2944
2945 Version 4.0 17-Feb-03
2946 ---------------------
2947
2948 1. If a comment in an extended regex that started immediately after a meta-item
2949 extended to the end of string, PCRE compiled incorrect data. This could lead to
2950 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2951
2952 2. Moved to autoconf 2.53 and libtool 1.4.2.
2953
2954 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2955 the special perltest8 script is no longer needed - all the tests can be run
2956 from a single perltest script.
2957
2958 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2959 by \s. It has now been removed in PCRE. This means it isn't recognized as
2960 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2961 class [:space:] *does* include VT, thereby creating a mess.
2962
2963 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2964 space and tab.
2965
2966 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2967 its new features into the main test script, reducing the number of scripts.
2968
2969 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2970 were backward compatible, and made the (?i) apply to the whole pattern, as if
2971 /i were given. Now it behaves more logically, and applies the option setting
2972 only to what follows. PCRE has been changed to follow suit. However, if it
2973 finds options settings right at the start of the pattern, it extracts them into
2974 the global options, as before. Thus, they show up in the info data.
2975
2976 8. Added support for the \Q...\E escape sequence. Characters in between are
2977 treated as literals. This is slightly different from Perl in that $ and @ are
2978 also handled as literals inside the quotes. In Perl, they will cause variable
2979 interpolation. Note the following examples:
2980
2981 Pattern PCRE matches Perl matches
2982
2983 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2984 \Qabc\$xyz\E abc\$xyz abc\$xyz
2985 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2986
2987 For compatibility with Perl, \Q...\E sequences are recognized inside character
2988 classes as well as outside them.
2989
2990 9. Re-organized 3 code statements in pcretest to avoid "overflow in
2991 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2992 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2993 signed/unsigned warnings.
2994
2995 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2996 option for pcretest, so I've replaced it by a simple function that does just
2997 that job.
2998
2999 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3000 "pcregrep -".
3001
3002 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3003 Java package. This provides some syntactic sugar for simple cases of what my
3004 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3005 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3006 item, you can use this simplified notation. Note that only makes sense with
3007 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3008 greediness, whatever the setting of the PCRE_UNGREEDY option.
3009
3010 13. A change of greediness default within a pattern was not taking effect at
3011 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3012 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3013 was abstracted outside.
3014
3015 14. PCRE now supports the \G assertion. It is true when the current matching
3016 position is at the start point of the match. This differs from \A when the
3017 starting offset is non-zero. Used with the /g option of pcretest (or similar
3018 code), it works in the same way as it does for Perl's /g option. If all
3019 alternatives of a regex begin with \G, the expression is anchored to the start
3020 match position, and the "anchored" flag is set in the compiled expression.
3021
3022 15. Some bugs concerning the handling of certain option changes within patterns
3023 have been fixed. These applied to options other than (?ims). For example,
3024 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3025 been the other way round. Some of this was related to change 7 above.
3026
3027 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3028 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3029 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3030 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3031
3032 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3033 mode. Unlike ".", it always matches newline, whatever the setting of
3034 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3035 assertions. Perl allows it, but it doesn't (in general) work because it can't
3036 calculate the length of the lookbehind. At least, that's the case for Perl
3037 5.8.0 - I've been told they are going to document that it doesn't work in
3038 future.
3039
3040 18. Added an error diagnosis for escapes that PCRE does not support: these are
3041 \L, \l, \N, \P, \p, \U, \u, and \X.
3042
3043 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3044 reading past the end of the pattern in cases such as /[abcd/.
3045
3046 20. PCRE was getting more memory than necessary for patterns with classes that
3047 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3048
3049 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3050 compiling PCRE for use with Virtual Pascal.
3051
3052 22. Small fix to the Makefile to make it work properly if the build is done
3053 outside the source tree.
3054
3055 23. Added a new extension: a condition to go with recursion. If a conditional
3056 subpattern starts with (?(R) the "true" branch is used if recursion has
3057 happened, whereas the "false" branch is used only at the top level.
3058
3059 24. When there was a very long string of literal characters (over 255 bytes
3060 without UTF support, over 250 bytes with UTF support), the computation of how
3061 much memory was required could be incorrect, leading to segfaults or other
3062 strange effects.
3063
3064 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3065 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3066 there was a subsequent back reference to those brackets. This meant that, for
3067 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3068 possible to check for precisely this case. All we can do is abandon the
3069 optimization if .* occurs inside capturing brackets when there are any back
3070 references whatsoever. (See below for a better fix that came later.)
3071
3072 26. The handling of the optimization for finding the first character of a
3073 non-anchored pattern, and for finding a character that is required later in the
3074 match were failing in some cases. This didn't break the matching; it just
3075 failed to optimize when it could. The way this is done has been re-implemented.
3076
3077 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3078
3079 28. Added a new feature that provides some of the functionality that Perl
3080 provides with (?{...}). The facility is termed a "callout". The way it is done
3081 in PCRE is for the caller to provide an optional function, by setting
3082 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3083 global variable. By default it is unset, which disables all calling out. To get
3084 the function called, the regex must include (?C) at appropriate points. This
3085 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3086 This provides a means of identifying different callout points. When PCRE
3087 reaches such a point in the regex, if pcre_callout has been set, the external
3088 function is called. It is provided with data in a structure called
3089 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3090 matching continues; if it returns a non-zero value, the match at the current
3091 point fails. However, backtracking will occur if possible. [This was changed
3092 later and other features added - see item 49 below.]
3093
3094 29. pcretest is upgraded to test the callout functionality. It provides a
3095 callout function that displays information. By default, it shows the start of
3096 the match and the current position in the text. There are some new data escapes
3097 to vary what happens:
3098
3099 \C+ in addition, show current contents of captured substrings
3100 \C- do not supply a callout function
3101 \C!n return 1 when callout number n is reached
3102 \C!n!m return 1 when callout number n is reached for the mth time
3103
3104 30. If pcregrep was called with the -l option and just a single file name, it
3105 output "<stdin>" if a match was found, instead of the file name.
3106
3107 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3108 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3109 pcre_exec(). This saves a malloc/free per call. The default value of
3110 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3111 when configuring.
3112
3113 32. The default maximum size of a compiled pattern is 64K. There have been a
3114 few cases of people hitting this limit. The code now uses macros to handle the
3115 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3116 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3117 configuring. Tests 2 and 5 work only with 2-byte links because they output
3118 debugging information about compiled patterns.
3119
3120 33. Internal code re-arrangements:
3121
3122 (a) Moved the debugging function for printing out a compiled regex into
3123 its own source file (printint.c) and used #include to pull it into
3124 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3125 separate copies.
3126
3127 (b) Defined the list of op-code names for debugging as a macro in
3128 internal.h so that it is next to the definition of the opcodes.
3129
3130 (c) Defined a table of op-code lengths for simpler skipping along compiled
3131 code. This is again a macro in internal.h so that it is next to the
3132 definition of the opcodes.
3133
3134 34. Added support for recursive calls to individual subpatterns, along the
3135 lines of Robin Houston's patch (but implemented somewhat differently).
3136
3137 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3138 allow it to read and process whole directories in Win32. This code was
3139 contributed by Lionel Fourquaux; it has not been tested by me.
3140
3141 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3142 used to name a group. Names consist of alphanumerics and underscores, and must
3143 be unique. Back references use the syntax (?P=name) and recursive calls use
3144 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3145 numbers. The function pcre_fullinfo() can be used after compilation to extract
3146 a name/number map. There are three relevant calls:
3147
3148 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3149 PCRE_INFO_NAMECOUNT yields the number of entries
3150 PCRE_INFO_NAMETABLE yields a pointer to the map.
3151
3152 The map is a vector of fixed-size entries. The size of each entry depends on
3153 the length of the longest name used. The first two bytes of each entry are the
3154 group number, most significant byte first. There follows the corresponding
3155 name, zero terminated. The names are in alphabetical order.
3156
3157 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3158 case instead of 255. Making it the same both with and without UTF-8 support
3159 means that the same test output works with both.
3160
3161 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3162 calling malloc() with a zero argument.
3163
3164 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3165 optimization. I've improved things by keeping a bitmap of backreferences with
3166 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3167 fact referenced, the optimization can be applied. It is unlikely that a
3168 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3169 the match to follow \n) will appear inside brackets with a number greater than
3170 31, but if it does, any back reference > 31 suppresses the optimization.
3171
3172 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3173 of disabling numbered capturing parentheses. Any opening parenthesis that is
3174 not followed by ? behaves as if it were followed by ?: but named parentheses
3175 can still be used for capturing (and they will acquire numbers in the usual
3176 way).
3177
3178 41. Redesigned the return codes from the match() function into yes/no/error so
3179 that errors can be passed back from deep inside the nested calls. A malloc
3180 failure while inside a recursive subpattern call now causes the
3181 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3182
3183 42. It is now possible to set a limit on the number of times the match()
3184 function is called in a call to pcre_exec(). This facility makes it possible to
3185 limit the amount of recursion and backtracking, though not in a directly
3186 obvious way, because the match() function is used in a number of different
3187 circumstances. The count starts from zero for each position in the subject
3188 string (for non-anchored patterns). The default limit is, for compatibility, a
3189 large number, namely 10 000 000. You can change this in two ways:
3190
3191 (a) When configuring PCRE before making, you can use --with-match-limit=n
3192 to set a default value for the compiled library.
3193
3194 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3195 a different value is set. See 45 below.
3196
3197 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3198
3199 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3200 of things that can be changed at compile time. The first argument specifies
3201 what is wanted and the second points to where the information is to be placed.
3202 The current list of available information is:
3203
3204 PCRE_CONFIG_UTF8
3205
3206 The output is an integer that is set to one if UTF-8 support is available;
3207 otherwise it is set to zero.
3208
3209 PCRE_CONFIG_NEWLINE
3210
3211 The output is an integer that it set to the value of the code that is used for
3212 newline. It is either LF (10) or CR (13).
3213
3214 PCRE_CONFIG_LINK_SIZE
3215
3216 The output is an integer that contains the number of bytes used for internal
3217 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3218
3219 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3220
3221 The output is an integer that contains the threshold above which the POSIX
3222 interface uses malloc() for output vectors. See item 31 above.
3223
3224 PCRE_CONFIG_MATCH_LIMIT
3225
3226 The output is an unsigned integer that contains the default limit of the number
3227 of match() calls in a pcre_exec() execution. See 42 above.
3228
3229 44. pcretest has been upgraded by the addition of the -C option. This causes it
3230 to extract all the available output from the new pcre_config() function, and to
3231 output it. The program then exits immediately.
3232
3233 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3234 order to support additional features. One way would have been to define
3235 pcre_exec2() (for example) with extra arguments, but this would not have been
3236 extensible, and would also have required all calls to the original function to
3237 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3238 is used for passing in "extra" data from pcre_study().
3239
3240 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3241 contains the following fields:
3242
3243 flags a bitmap indicating which of the following fields are set
3244 study_data opaque data from pcre_study()
3245 match_limit a way of specifying a limit on match() calls for a specific
3246 call to pcre_exec()
3247 callout_data data for callouts (see 49 below)
3248
3249 The flag bits are also defined in pcre.h, and are
3250
3251 PCRE_EXTRA_STUDY_DATA
3252 PCRE_EXTRA_MATCH_LIMIT
3253 PCRE_EXTRA_CALLOUT_DATA
3254
3255 The pcre_study() function now returns one of these new pcre_extra blocks, with
3256 the actual study data pointed to by the study_data field, and the
3257 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3258 before. That is, this change is entirely upwards-compatible and requires no
3259 change to existing code.
3260
3261 If you want to pass in additional data to pcre_exec(), you can either place it
3262 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3263 block.
3264
3265 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3266 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3267 times with different match limits, until it finds the minimum value needed for
3268 pcre_exec() to complete. The value is then output. This can be instructive; for
3269 most simple matches the number is quite small, but for pathological cases it
3270 gets very large very quickly.
3271
3272 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3273 returns the size of the data block pointed to by the study_data field in a
3274 pcre_extra block, that is, the value that was passed as the argument to
3275 pcre_malloc() when PCRE was getting memory in which to place the information
3276 created by pcre_study(). The fourth argument should point to a size_t variable.
3277 pcretest has been extended so that this information is shown after a successful
3278 pcre_study() call when information about the compiled regex is being displayed.
3279
3280 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3281 because what follows is always an absolute path. (Later: it turns out that this
3282 is more than cosmetic for MinGW, because it doesn't like empty path
3283 components.)
3284
3285 49. Some changes have been made to the callout feature (see 28 above):
3286
3287 (i) A callout function now has three choices for what it returns:
3288
3289 0 => success, carry on matching
3290 > 0 => failure at this point, but backtrack if possible
3291 < 0 => serious error, return this value from pcre_exec()
3292
3293 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3294 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3295 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3296 use by callout functions. It will never be used by PCRE itself.
3297
3298 (ii) The pcre_extra structure (see 45 above) has a void * field called
3299 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3300 pcre_callout_block structure has a field of the same name. The contents of
3301 the field passed in the pcre_extra structure are passed to the callout
3302 function in the corresponding field in the callout block. This makes it
3303 easier to use the same callout-containing regex from multiple threads. For
3304 testing, the pcretest program has a new data escape
3305
3306 \C*n pass the number n (may be negative) as callout_data
3307
3308 If the callout function in pcretest receives a non-zero value as
3309 callout_data, it returns that value.
3310
3311 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3312 there were some redundant $(CFLAGS) in commands that are now specified as
3313 $(LINK), which already includes $(CFLAGS).
3314
3315 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3316 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3317 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3318 one-byte characters throughout. Note that case-insensitive matching applies
3319 only to characters whose values are less than 256. PCRE doesn't support the
3320 notion of cases for higher-valued characters.
3321
3322 (i) A character class whose characters are all within 0-255 is handled as
3323 a bit map, and the map is inverted for negative classes. Previously, a
3324 character > 255 always failed to match such a class; however it should
3325 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3326
3327 (ii) A negated character class with a single character < 255 is coded as
3328 "not this character" (OP_NOT). This wasn't working properly when the test
3329 character was multibyte, either singly or repeated.
3330
3331 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3332 mode, for example: \x{100}{2,3}.
3333
3334 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3335 singly or repeated) now correctly test multibyte characters. However,
3336 PCRE doesn't recognize any characters with values greater than 255 as
3337 digits, spaces, or word characters. Such characters always match \D, \S,
3338 and \W, and never match \d, \s, or \w.
3339
3340 (v) Classes may now contain characters and character ranges with values
3341 greater than 255. For example: [ab\x{100}-\x{400}].
3342
3343 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3344 PCRE in UTF-8 mode.
3345
3346 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3347 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3348 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3349 value.)
3350
3351 53. The single man page has become too large. I have therefore split it up into
3352 a number of separate man pages. These also give rise to individual HTML pages;
3353 these are now put in a separate directory, and there is an index.html page that
3354 lists them all. Some hyperlinking between the pages has been installed.
3355
3356 54. Added convenience functions for handling named capturing parentheses.
3357
3358 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3359 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3360 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3361 are faulted.
3362
3363 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3364 calling configure. These values are used when compiling the dftables.c program
3365 which is run to generate the source of the default character tables. They
3366 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3367 you will need to set these values.
3368
3369 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3370
3371
3372 Version 3.9 02-Jan-02
3373 ---------------------
3374
3375 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3376
3377 2. If --disable-static was given, the building process failed when trying to
3378 build pcretest and pcregrep. (For some reason it was using libtool to compile
3379 them, which is not right, as they aren't part of the library.)
3380
3381
3382 Version 3.8 18-Dec-01
3383 ---------------------
3384
3385 1. The experimental UTF-8 code was completely screwed up. It was packing the
3386 bytes in the wrong order. How dumb can you get?
3387
3388
3389 Version 3.7 29-Oct-01
3390 ---------------------
3391
3392 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3393 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3394 this didn't happen under Solaris 8, where I normally test things.
3395
3396 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3397 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3398 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3399 this fix an hour or so after the initial 3.7 release.)
3400
3401
3402 Version 3.6 23-Oct-01
3403 ---------------------
3404
3405 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3406 offsets passed as NULL with zero offset count.
3407
3408 2. The config.guess and config.sub files had not been updated when I moved to
3409 the latest autoconf.
3410
3411
3412 Version 3.5 15-Aug-01
3413 ---------------------
3414
3415 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3416 had been forgotten.
3417
3418 2. By using declared but undefined structures, we can avoid using "void"
3419 definitions in pcre.h while keeping the internal definitions of the structures
3420 private.
3421
3422 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3423 user point of view, this means that both static and shared libraries are built
3424 by default, but this can be individually controlled. More of the work of
3425 handling this static/shared cases is now inside libtool instead of PCRE's make
3426 file.
3427
3428 4. The pcretest utility is now installed along with pcregrep because it is
3429 useful for users (to test regexs) and by doing this, it automatically gets
3430 relinked by libtool. The documentation has been turned into a man page, so
3431 there are now .1, .txt, and .html versions in /doc.
3432
3433 5. Upgrades to pcregrep:
3434 (i) Added long-form option names like gnu grep.
3435 (ii) Added --help to list all options with an explanatory phrase.
3436 (iii) Added -r, --recursive to recurse into sub-directories.
3437 (iv) Added -f, --file to read patterns from a file.
3438
3439 6. pcre_exec() was referring to its "code" argument before testing that
3440 argument for NULL (and giving an error if it was NULL).
3441
3442 7. Upgraded Makefile.in to allow for compiling in a different directory from
3443 the source directory.
3444
3445 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3446 options bits, the pointer it was passed was to an int instead of to an unsigned
3447 long int. This mattered only on 64-bit systems.
3448
3449 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3450 generated) instead of pcre.in, which it its source. Also made the same change
3451 in several of the .c files.
3452
3453 10. A new release of gcc defines printf() as a macro, which broke pcretest
3454 because it had an ifdef in the middle of a string argument for printf(). Fixed
3455 by using separate calls to printf().
3456
3457 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3458 script, to force use of CR or LF instead of \n in the source. On non-Unix
3459 systems, the value can be set in config.h.
3460
3461 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3462 absolute limit. Changed the text of the error message to make this clear, and
3463 likewise updated the man page.
3464
3465 13. The limit of 99 on the number of capturing subpatterns has been removed.
3466 The new limit is 65535, which I hope will not be a "real" limit.
3467
3468
3469 Version 3.4 22-Aug-00
3470 ---------------------
3471
3472 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3473
3474 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3475
3476
3477 Version 3.3 01-Aug-00
3478 ---------------------
3479
3480 1. If an octal character was given, but the value was greater than \377, it
3481 was not getting masked to the least significant bits, as documented. This could
3482 lead to crashes in some systems.
3483
3484 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3485 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3486
3487 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3488 These just pass their arguments on to (pcre_free)(), but they are provided
3489 because some uses of PCRE bind it to non-C systems that can call its functions,
3490 but cannot call free() or pcre_free() directly.
3491
3492 4. Add "make test" as a synonym for "make check". Corrected some comments in
3493 the Makefile.
3494
3495 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3496 Makefile.
3497
3498 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3499 command called pgrep for grepping around the active processes.
3500
3501 7. Added the beginnings of support for UTF-8 character strings.
3502
3503 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3504 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3505 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3506 out for the ar command.)
3507
3508
3509 Version 3.2 12-May-00
3510 ---------------------
3511
3512 This is purely a bug fixing release.
3513
3514 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3515 of ZA. This was just one example of several cases that could provoke this bug,
3516 which was introduced by change 9 of version 2.00. The code for breaking
3517 infinite loops after an iteration that matches an empty string was't working
3518 correctly.
3519
3520 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3521 when matched against abbab (for example). After matching an empty string, it
3522 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3523 caused it to match further down the string than it should.
3524
3525 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3526 was there because it doesn't seem to be needed, and it causes trouble on some
3527 systems, as it is not a Standard C header. It has been removed.
3528
3529 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3530 were reported on the Macintosh. The changes were from
3531
3532 while ((c = *(++ptr)) != 0 && c != '\n');
3533 to
3534 while ((c = *(++ptr)) != 0 && c != '\n') ;
3535
3536 Totally extraordinary, but if that's what it takes...
3537
3538 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3539 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3540 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3541 assumes the way PCRE uses memmove() (always moving upwards).
3542
3543 6. PCRE is being used in one environment where strchr() is not available. There
3544 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3545 faster code anyway.
3546
3547
3548 Version 3.1 09-Feb-00
3549 ---------------------
3550
3551 The only change in this release is the fixing of some bugs in Makefile.in for
3552 the "install" target:
3553
3554 (1) It was failing to install pcreposix.h.
3555
3556 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3557
3558
3559 Version 3.0 01-Feb-00
3560 ---------------------
3561
3562 1. Add support for the /+ modifier to perltest (to output $` like it does in
3563 pcretest).
3564
3565 2. Add support for the /g modifier to perltest.
3566
3567 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3568 matches null strings.
3569
3570 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3571 pattern. Perl treats empty patterns specially - it reuses the most recent
3572 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3573 effect.
3574
3575 5. The POSIX interface was broken in that it was just handing over the POSIX
3576 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3577 required a bigger vector, with some working space on the end. This means that
3578 the POSIX wrapper now has to get and free some memory, and copy the results.
3579
3580 6. Added some simple autoconf support, placing the test data and the
3581 documentation in separate directories, re-organizing some of the
3582 information files, and making it build pcre-config (a GNU standard). Also added
3583 libtool support for building PCRE as a shared library, which is now the
3584 default.
3585
3586 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3587 09 are not valid octal constants. Single digits will be used for minor values
3588 less than 10.
3589
3590 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3591 existing programs that set these in the POSIX interface can use PCRE without
3592 modification.
3593
3594 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3595 return all that pcre_info() returns, plus additional data. The pcre_info()
3596 function is retained for compatibility, but is considered to be obsolete.
3597
3598 10. Added experimental recursion feature (?R) to handle one common case that
3599 Perl 5.6 will be able to do with (?p{...}).
3600
3601 11. Added support for POSIX character classes like [:alpha:], which Perl is
3602 adopting.
3603
3604
3605 Version 2.08 31-Aug-99
3606 ----------------------
3607
3608 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3609 trying to match at the startoffset position, but instead was moving forward to
3610 the next newline as if a previous match had failed.
3611
3612 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3613 and could get into a loop if a null string was matched other than at the start
3614 of the subject.
3615
3616 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3617 be distinguished at compile time, and for completeness also added PCRE_DATE.
3618
3619 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3620 in GnuWin32 environments.
3621
3622
3623 Version 2.07 29-Jul-99
3624 ----------------------
3625
3626 1. The documentation is now supplied in plain text form and HTML as well as in
3627 the form of man page sources.
3628
3629 2. C++ compilers don't like assigning (void *) values to other pointer types.
3630 In particular this affects malloc(). Although there is no problem in Standard
3631 C, I've put in casts to keep C++ compilers happy.
3632
3633 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3634 should be (const char *).
3635
3636 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3637 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3638 However, I haven't made this a standard facility. The documentation doesn't
3639 mention it, and the Makefile doesn't support it.
3640
3641 5. The Makefile now contains an "install" target, with editable destinations at
3642 the top of the file. The pcretest program is not installed.
3643
3644 6. pgrep -V now gives the PCRE version number and date.
3645
3646 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3647 causing the entire string to be ignored, instead of just the last character.
3648
3649 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3650 non-matching string, it can take a very, very long time, even for strings of
3651 quite modest length, because of the nested recursion. PCRE now does better in
3652 some of these cases. It does this by remembering the last required literal
3653 character in the pattern, and pre-searching the subject to ensure it is present
3654 before running the real match. In other words, it applies a heuristic to detect
3655 some types of certain failure quickly, and in the above example, if presented
3656 with a string that has no trailing " it gives "no match" very quickly.
3657
3658 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3659 other alternatives are tried instead.
3660
3661
3662 Version 2.06 09-Jun-99
3663 ----------------------
3664
3665 1. Change pcretest's output for amount of store used to show just the code
3666 space, because the remainder (the data block) varies in size between 32-bit and
3667 64-bit systems.
3668
3669 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3670 start matching at. This allows lookbehinds to work when searching for multiple
3671 occurrences in a string.
3672
3673 3. Added additional options to pcretest for testing multiple occurrences:
3674
3675 /+ outputs the rest of the string that follows a match
3676 /g loops for multiple occurrences, using the new startoffset argument
3677 /G loops for multiple occurrences by passing an incremented pointer
3678
3679 4. PCRE wasn't doing the "first character" optimization for patterns starting
3680 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3681 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3682 the letter 'x'. On long subject strings, this gives a significant speed-up.
3683
3684
3685 Version 2.05 21-Apr-99
3686 ----------------------
3687
3688 1. Changed the type of magic_number from int to long int so that it works
3689 properly on 16-bit systems.
3690
3691 2. Fixed a bug which caused patterns starting with .* not to work correctly
3692 when the subject string contained newline characters. PCRE was assuming
3693 anchoring for such patterns in all cases, which is not correct because .* will
3694 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3695 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3696 must be retried after every newline in the subject.
3697
3698
3699 Version 2.04 18-Feb-99
3700 ----------------------
3701
3702 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3703 computation of the store needed to hold the pattern was incorrect (too large).
3704 If such patterns were nested a few deep, this could multiply and become a real
3705 problem.
3706
3707 2. Added /M option to pcretest to show the memory requirement of a specific
3708 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3709
3710 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3711 compiled in such a way that the backtracking after subsequent failure was
3712 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3713 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3714
3715
3716 Version 2.03 02-Feb-99
3717 ----------------------
3718
3719 1. Fixed typo and small mistake in man page.
3720
3721 2. Added 4th condition (GPL supersedes if conflict) and created separate
3722 LICENCE file containing the conditions.
3723
3724 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3725 Perl, that is the internal \ allows the delimiter to be included in the
3726 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3727 the final delimiter, add \ to the end of the pattern (to test the error).
3728
3729 4. Added the convenience functions for extracting substrings after a successful
3730 match. Updated pcretest to make it able to test these functions.
3731
3732
3733 Version 2.02 14-Jan-99
3734 ----------------------
3735
3736 1. Initialized the working variables associated with each extraction so that
3737 their saving and restoring doesn't refer to uninitialized store.
3738
3739 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3740 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3741 fix the problem.
3742
3743 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3744 calls, and wasn't printing the correct value for compiling calls. Increased the
3745 default value of LOOPREPEAT, and the number of significant figures in the
3746 times.
3747
3748 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3749
3750 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3751 a building problem on Windows NT with a FAT file system.
3752
3753
3754 Version 2.01 21-Oct-98
3755 ----------------------
3756
3757 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3758 to character tables built by pcre_maketables() in the current locale. If NULL
3759 is passed, the default tables are used.
3760
3761
3762 Version 2.00 24-Sep-98
3763 ----------------------
3764
3765 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3766 it any more.
3767
3768 2. Allow quantification of (?>) groups, and make it work correctly.
3769
3770 3. The first character computation wasn't working for (?>) groups.
3771
3772 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3773 end of the subject) and add 5.005's \z, which really does match only at the
3774 very end of the subject.
3775
3776 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3777
3778 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3779 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3780 localized options. All options to pcre_study() were also removed.
3781
3782 7. Add other new features from 5.005:
3783
3784 $(?<= positive lookbehind
3785 $(?<! negative lookbehind
3786 (?imsx-imsx) added the unsetting capability
3787 such a setting is global if at outer level; local otherwise
3788 (?imsx-imsx:) non-capturing groups with option setting
3789 (?(cond)re|re) conditional pattern matching
3790
3791 A backreference to itself in a repeated group matches the previous
3792 captured string.
3793
3794 8. General tidying up of studying (both automatic and via "study")
3795 consequential on the addition of new assertions.
3796
3797 9. As in 5.005, unlimited repeated groups that could match an empty substring
3798 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3799 runtime if any iteration does actually match an empty substring.
3800
3801 10. Include the RunTest script in the distribution.
3802
3803 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3804 discrepancies, some of which were old and were also with respect to 5.004. They
3805 have now been fixed.
3806
3807
3808 Version 1.09 28-Apr-98
3809 ----------------------
3810
3811 1. A negated single character class followed by a quantifier with a minimum
3812 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3813 program crashes, or just wrong answers. This did not apply to negated classes
3814 containing more than one character, or to minima other than one.
3815
3816
3817 Version 1.08 27-Mar-98
3818 ----------------------
3819
3820 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3821
3822 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3823 latter must appear before anything that relies on it in the pattern.
3824
3825
3826 Version 1.07 16-Feb-98
3827 ----------------------
3828
3829 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3830 repeat of a potentially empty string).
3831
3832
3833 Version 1.06 23-Jan-98
3834 ----------------------
3835
3836 1. Added Markus Oberhumer's little patches for C++.
3837
3838 2. Literal strings longer than 255 characters were broken.
3839
3840
3841 Version 1.05 23-Dec-97
3842 ----------------------
3843
3844 1. Negated character classes containing more than one character were failing if
3845 PCRE_CASELESS was set at run time.
3846
3847
3848 Version 1.04 19-Dec-97
3849 ----------------------
3850
3851 1. Corrected the man page, where some "const" qualifiers had been omitted.
3852
3853 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3854 input syntax.
3855
3856 3. Fixed memory leak which occurred when a regex with back references was
3857 matched with an offsets vector that wasn't big enough. The temporary memory
3858 that is used in this case wasn't being freed if the match failed.
3859
3860 4. Tidied pcretest to ensure it frees memory that it gets.
3861
3862 5. Temporary memory was being obtained in the case where the passed offsets
3863 vector was exactly big enough.
3864
3865 6. Corrected definition of offsetof() from change 5 below.
3866
3867 7. I had screwed up change 6 below and broken the rules for the use of
3868 setjmp(). Now fixed.
3869
3870
3871 Version 1.03 18-Dec-97
3872 ----------------------
3873
3874 1. A erroneous regex with a missing opening parenthesis was correctly
3875 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3876 on some systems.
3877
3878 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3879 it was reported that one broken compiler failed on the former because "code" is
3880 also an independent variable.
3881
3882 3. The erroneous regex a[]b caused an array overrun reference.
3883
3884 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3885 fail on data ending with that character. (It was going on too far, and checking
3886 the next character, typically a binary zero.) This was specific to the
3887 optimized code for single-character negative classes.
3888
3889 5. Added a contributed patch from the TIN world which does the following:
3890
3891 + Add an undef for memmove, in case the the system defines a macro for it.
3892
3893 + Add a definition of offsetof(), in case there isn't one. (I don't know
3894 the reason behind this - offsetof() is part of the ANSI standard - but
3895 it does no harm).
3896
3897 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3898 most of the places where whitespace preceded '#'. I have given up and
3899 allowed the remaining 2 cases to be at the margin.
3900
3901 + Rename some variables in pcre to eliminate shadowing. This seems very
3902 pedantic, but does no harm, of course.
3903
3904 6. Moved the call to setjmp() into its own function, to get rid of warnings
3905 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3906
3907 7. Constructs such as \d{8,} were compiling into the equivalent of
3908 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3909 outcome, but in this particular case used more store than had been allocated,
3910 which caused the bug to be discovered because it threw up an internal error.
3911
3912 8. The debugging code in both pcre and pcretest for outputting the compiled
3913 form of a regex was going wrong in the case of back references followed by
3914 curly-bracketed repeats.
3915
3916
3917 Version 1.02 12-Dec-97
3918 ----------------------
3919
3920 1. Typos in pcre.3 and comments in the source fixed.
3921
3922 2. Applied a contributed patch to get rid of places where it used to remove
3923 'const' from variables, and fixed some signed/unsigned and uninitialized
3924 variable warnings.
3925
3926 3. Added the "runtest" target to Makefile.
3927
3928 4. Set default compiler flag to -O2 rather than just -O.
3929
3930
3931 Version 1.01 19-Nov-97
3932 ----------------------
3933
3934 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3935 like /([ab]*)*/, that is, for classes with more than one character in them.
3936
3937 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3938 as /((?>a*))*/ (a PCRE_EXTRA facility).
3939
3940
3941 Version 1.00 18-Nov-97
3942 ----------------------
3943
3944 1. Added compile-time macros to support systems such as SunOS4 which don't have
3945 memmove() or strerror() but have other things that can be used instead.
3946
3947 2. Arranged that "make clean" removes the executables.
3948
3949
3950 Version 0.99 27-Oct-97
3951 ----------------------
3952
3953 1. Fixed bug in code for optimizing classes with only one character. It was
3954 initializing a 32-byte map regardless, which could cause it to run off the end
3955 of the memory it had got.
3956
3957 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3958
3959
3960 Version 0.98 22-Oct-97
3961 ----------------------
3962
3963 1. Fixed bug in code for handling temporary memory usage when there are more
3964 back references than supplied space in the ovector. This could cause segfaults.
3965
3966
3967 Version 0.97 21-Oct-97
3968 ----------------------
3969
3970 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3971
3972 2. Optimized negated single characters not to use a bit map.
3973
3974 3. Brought error texts together as macro definitions; clarified some of them;
3975 fixed one that was wrong - it said "range out of order" when it meant "invalid
3976 escape sequence".
3977
3978 4. Changed some char * arguments to const char *.
3979
3980 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3981
3982 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3983 pcretest.
3984
3985
3986 Version 0.96 16-Oct-97
3987 ----------------------
3988
3989 1. Added a simple "pgrep" utility to the distribution.
3990
3991 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3992 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3993 where "ddd" means "one or more decimal digits".
3994
3995 3. Fixed serious bug. If a pattern had a back reference, but the call to
3996 pcre_exec() didn't supply a large enough ovector to record the related
3997 identifying subpattern, the match always failed. PCRE now remembers the number
3998 of the largest back reference, and gets some temporary memory in which to save
3999 the offsets during matching if necessary, in order to ensure that
4000 backreferences always work.
4001
4002 4. Increased the compatibility with Perl in a number of ways:
4003
4004 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4005 to request this handling. The option can be set at compile or exec time.
4006
4007 (b) $ matches before a terminating newline by default; an option
4008 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4009 mode). The option can be set at compile or exec time.
4010
4011 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4012 the same as Perl's. If the decimal number it represents is less than 10
4013 or there aren't that many previous left capturing parentheses, an octal
4014 escape is read. Inside a character class, it's always an octal escape,
4015 even if it is a single digit.
4016
4017 (d) An escaped but undefined alphabetic character is taken as a literal,
4018 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4019 escapes.
4020
4021 (e) {0} is now permitted. (The previous item is removed from the compiled
4022 pattern).
4023
4024 5. Changed all the names of code files so that the basic parts are no longer
4025 than 10 characters, and abolished the teeny "globals.c" file.
4026
4027 6. Changed the handling of character classes; they are now done with a 32-byte
4028 bit map always.
4029
4030 7. Added the -d and /D options to pcretest to make it possible to look at the
4031 internals of compilation without having to recompile pcre.
4032
4033
4034 Version 0.95 23-Sep-97
4035 ----------------------
4036
4037 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4038 \x20 at the start of a run of normal characters. These were being treated as
4039 real characters, instead of the source characters being re-checked.
4040
4041
4042 Version 0.94 18-Sep-97
4043 ----------------------
4044
4045 1. The functions are now thread-safe, with the caveat that the global variables
4046 containing pointers to malloc() and free() or alternative functions are the
4047 same for all threads.
4048
4049 2. Get pcre_study() to generate a bitmap of initial characters for non-
4050 anchored patterns when this is possible, and use it if passed to pcre_exec().
4051
4052
4053 Version 0.93 15-Sep-97
4054 ----------------------
4055
4056 1. /(b)|(:+)/ was computing an incorrect first character.
4057
4058 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4059 but not actually doing anything yet.
4060
4061 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4062 as Perl does (e.g. [-az] or [az-]).
4063
4064 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4065 all possible positions.
4066
4067 5. Split up into different modules to avoid including unneeded functions in a
4068 compiled binary. However, compile and exec are still in one module. The "study"
4069 function is split off.
4070
4071 6. The character tables are now in a separate module whose source is generated
4072 by an auxiliary program - but can then be edited by hand if required. There are
4073 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4074 toupper() in the code.
4075
4076 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4077 make them global. Abolish the function for setting them, as the caller can now
4078 set them directly.
4079
4080
4081 Version 0.92 11-Sep-97
4082 ----------------------
4083
4084 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4085 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4086
4087 2. Caseless matching was not working in character classes if the characters in
4088 the pattern were in upper case.
4089
4090 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4091
4092 4. Make PCRE_ANCHORED public and accept as a compile option.
4093
4094 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4095 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4096 pass them.
4097
4098 6. Give an error if bad option bits passed at compile or run time.
4099
4100 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4101 pcretest to cause it to pass that flag.
4102
4103 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4104 options, and the first character, if set.
4105
4106 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4107
4108
4109 Version 0.91 10-Sep-97
4110 ----------------------
4111
4112 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4113 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4114
4115 2. PCRE was looping on encountering an indefinitely repeated back reference to
4116 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4117 Perl does - treats the match as successful.
4118
4119 ****

Properties

Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12